Sie abeiere! isin eee rive oe aban ors Museum of a, ca j Nag “97 % io) 4 1869 THE LIBRARY OCEANIC BIRDS OF SOUTH AMERICA y ’ a < " 7 > ty | “a ' D. r m eo %, q * i | ¥ wf AT ' nes - —. : i 4 ii, rs i ‘i i = OCEANIC BIRDS OF SOUTH AMERICA A STUDY OF SPECIES OF THE RELATED COASTS AND SEAS, INCLUDING THE AMERICAN QUADRANT OF ANTARCTICA BASED UPON THE BREWSTER - SANFORD COLLECTION IN THE AMERICAN MUSEUM OF NATURAL HISTORY BY ROBERT CUSHMAN ,MURPHY ILLUSTRATED FROM PAINTINGS BY FRANCIS L. JAQUES PHOTOGRAPHS, MAPS, AND OTHER DRAWINGS Vol. I THE MACMILLAN COMPANY THE AMERICAN MUSEUM OF NATURAL HISTORY NEW YORK COPYRIGHT 1936, BY THE AMERICAN MUSEUM OF NATURAL HISTORY To FRANK M. CHAPMAN OCEANIC BIRDS OF SOUTH AMERICA CONTENTS PART I. THE PHYSICAL ENVIRONMENT INTRODUCTION Tue Frztp-WorKER NARRATIVE OF THE BREWSTER-SANFORD EXPEDITION 1912 . 1913. 1914. 1915. 1916. 1917 . Oruer Fietp Work CoNCERNED WITH THE PRODUCTION OF THIs REPORT 1 South Georgia Expedition, R. C. Murphy 2. Correia’s Work at South Georgia ae) FOS myanaw . Cape Verde Island Expedition, J. G. Correia . . Peruvian Littoral Expedition, R. C. Murphy . The Visit of Francis L. Jaques to Peru . Ecuador and Northern Peru, Murphy and Heilner The Explorations of Chapman and His Associates Tate's Work on the Coast of Venezuela Chapin’s Visit to the Galapagos . The Cruise of the ‘Zaca’.. . Miscellaneous Sources of Material ACKNOWLEDGMENTS THE GEOGRAPHIC BACKGROUND . 1. The Continent . 2. Meteorology . Trade Wind Régimes : . Maritime Climatic Control o on the Central East Coast . Modification of the Double Equatorial Rainy Season The Northern-Summer Rainfall Region The Northwestern Coastal Area of Perpetual Rains . Winds and Seasons in Guiana The Line of Doldrums . The Southern-Summer Rainfall Region . The Amazonian Climatic District mw ROO mw Rin SA Vil vill CONTENTS The Chilean Mediterranean. . The Brazilian Semi-Arid Monsoon District . Rainfall on the Brazilian Coast North of Rio . . Anticyclonic Coasts and Transcontinental Lows The Zone of Westerlies . . The Diagonal Desert Strip . Sea Birds in Relation to Gales mad ice Shona . Birds and Hurricanes mito 8 Fw ew. Hurricanes of: (1) August, 1879 (2) August, 1893 (3) August, 1933. (4) September, 1876 (5) September, 1878 . . (6) September-October, 1898 Tue Hypro.ocy 1n RevaTIoNn To Oceanic BirDs . 1. The Nutritional Basis of Marine Life . 2. The Zones of Surface Water . a. The Antarctic Zone ; (1) Birds Typical of the Antatctic Zone b. The Sub-Antarctic Zone (1) Birds Characteristic of the Sub- Antarctic Zone (2) Birds Common to un the Antarctic and Sub-Ant- arctic Zones c. The Sub-Tropical Zone . d. The Tropical Zone... (1) Tropical and Sub- Tropical os Birds 3. Ocean Currents. . a. The Central Anticyclonic Regions b. The North Equatorial Current c. The Brazil Current . (1) The Poverty of Tropical Bird Life off Soundings . d. The Cape Horn Drift . The Falkland Current f. The Humboldt Current... (1) The Provenance of Humboldt Current Birds | (2) The Provenance of Galapagos Sea Birds a g. The Nifio Current hb. Meteorological, Guscneenaite and Biological Correlations in the Northwestern Bight of South America (1) Zoneless Birds CONTENTS ix Pace AN ORNITHOLOGICAL CIRCUMNAVIGATION OF SOUTH AMERICA. : a) GK, 1. The Caribbean Coast... Saeed yay «8 Tato 2. The Atlantic Coast, Trinidad to the Shoulder of Brazil rae (1) Reef Corals in South America eG a Sm ae a. whe Arlantic Equatorialdslands ..).4..).) t.cs 4 wes? JUS a. Rocas Reef. arse a er. ee eg b. Fernando Noronha ey So er eer ae Wee ee c. St. Paul Rocks a Ue WE Pe ae ee d. Ascension ee ee ee ee ee erg Sw AIS Cappeoiiets:; yh ro 8 Le Se ee ae ees 4. Recife de Pernambuco to the Riodela Plata. . . . . 4156 joe Ne, peantic’sub-Lropical\Islands . «. 4.4. a ay ee a. South Trinidad and Martin Vas ar ee ee (Ay ited saltiness aay te, Pat 2h) Se ko 6. Southern Coasts, Gulf of La Plata to the Strait of Dacettan os: Se eae (1) Sedentary and Exclusively Magellanic Salt-Water Birds : 198 (2) Magellanic Migratory Sale: Water His, : 199 G) Sedentary Salt-Water Birds of Combined Magel- lanic and Humboldt Current Range... 200 (4) Widely Distributed Sub-Antarctic Salt-Water Bade with South American Breeding Ranges Restricted to the Magellanic District... 200 (5) Magellanic Salt-Water Birds of Bon Anteeceie or More Extensive Breeding Range ait: 201 7: The Aclastic Sab-Aararctic Islands). 9 im,’ See S202 grepaiona islarids 1+. + a ” WW puch @ oy steele ake eesowebigland 230 b ee sR Ge ke ae poiememas nha ae. oo Nw 5) aa a eee 8. The Scotia Arc ~ Wm, 5 eee, a eee a. South Georgia. ee ee eee b. South Sandwich Group n'y ER” pautdy i pacthe Gayton Py ome GRIEVE 8 toh ae of ie foes ek” ae oa da. South Shetlands... 20 RU hg) cog el ere N e. The Antarctic Archipelago . ORE MR a A (2bouver Isiagd, .. 4/00" A ee ee 9: Hastetn Puegia to.CentraliChile $2 wg ee ag a. Staten Island . «Wi eee sc ee ee b. Tierra del Fuego and Patagonian Chile” Ge catenin ta 235 c. The Gulf of Coronados to Valparaiso . . . . . 249 10. ae 12" 113), 14. CONTENTS The Pacific Sub-Tropical Islands a. Juan Fernandez b. San Ambrosio and San Felix (1) Polynesian Sources of West Coast Wonieee The Desert Coast—Coquimbo to Point Parifias a. Guano The Galapagos Archipelago Northern Peru to Panama Islands of the Tropical Pacific oe a. Cocos Island . b. Malpelo Island c. The Pearl Islands . PART II. THE OCEANIC BIRDS Scope and Method of Treatment Forms Described as New in This Book ORDER SPHENISCIFORMES Tue Pencurns, Family Spheniscidae Ancestry : Morphological a Physiolowedh falecrelanionchins Psychobiologic Reactions . Distribution . King Penguin Aprenodytes patagonicus J. F. Miller . ‘Emperor Penguin Aptenodytes forsteri G. R. Gray Gentoo Penguin Pygoscelis papua (Forster) Adélie Penguin Pygoscelis adeliae (Hombron and jac Ringed Penguin Pygoscelis antarctica (Forster ) Rockhopper Penguin Exdyptes crestatus (J. F. Miller) Macaroni Penguin Eudyptes chrysolophus (Brandt) . Magellanic Penguin Spheniscus magellanicus (Forster) Peruvian Penguin Spheniscus humboldti Meyen Galapagos Penguin Spheniscus mendiculus Sundevall Pace 254 254 258 262 263 286 296 303 317 317 319 320 323 328 329 323 330 334 338 342 343 354 367 386 406 415 432 437 452 466 CONTENTS x1 Pace ORDER PROCELLARUIFOR MES: ple ¢ ic! Wits Ge een! a a) oe AAD Classification Poe Ihe, POL SOP A eae, | Sty rik PATE cicolree felch.) a ee es ere at hh aad a Dist peciOal et Ps) ya. ne ote est Weed os, tase ae. Ree Evolution... Nike yh ectle are a Vices edu sete se pean Nidification and Granth ae Ce mete ny ae ee ree BOOMERS © fc) Tk ites Bae. on BOON wel ae Oe eee eee ae DCOMONIONS & jie. tly ae Sg iets WM er jagel lt 3) ae i Walletgigin’ erage ae ee he le toes Sia Dae Enns CO As Mk tN INET Shek AVES edie Seg ghia beth ere BCCCLIG a: Xm eugeld Nn gee ame) ea Mane #) aiuh teeing bee § a Tue Acpatrosses, Family Diomedeidae . . . . . = «. 489 Sooty Albatross Phoebetria fusca (Hilsenberg). Swe 494 Light-mantled Sooty Albatross Phoebetria palpebrata (Forster) 3 UP ce eo Black-browed Albatross Diomedea melanophris Temminck . . . . 505 Gray-headed Albatross Diomedea chrysostoma Forster . . . . . S514 Yellow-nosed Albatross Diomedea chlororhynchosGmelin . . . . S518 Buller’s Albatross Diomedea bullert Rothschild 0 Meter a. White-capped Albatross Diomedea cauta salvini (Rothschild) ak ieeeeo Galapagos Albatross Diomedea irrorata Salvin... eed eee Wandering Albatross Diomedea exulans exulans Linnaeus ot , Tristan Wandering Albatross Diomedea exulans dabbenena Mathews Tp tee) toe Royal Albatross Diomedea epomophora Lesson. . . . . «~~ «575 Tue Petrets, Family Procellariidae . . 584 The Fulmars, Cape Pigeons, Whale-birds, Galt panels sicae waters, etc., Subfamily Puffininae .. . 584 Giant Fulmar Macronectes giganteus(Gmelin). . . . . . . 584 Silver-gray Fulmar Priocella antarctica (Stephens) . . . . . . 596 Cape Pigeon Daption capensis (Linnaeus) 5 ets alia ae a) eee a OO The Whale-birds of the Genus Pachyptila : te MEE rey rors! Broad-billed Whale-bird Pachyprila forsteri (Latham) ee OLD Olsy Antarctic Whale-bird Pachyptila desolata(Gmelin) . . . . 613,620 Slender-billed Whale-bird Pachyprila belcheri alae Se Gia ye Fairy Prion Pachyptila turtur (Kuhl) : es a oer Snow Petrel Pagodroma nivea (Forster). eine ace bed oe oe Antarctic Petrel Thalassoica antarctica (Gmelin) PR ere Semen eco: Shoemaker Procellaria aequinoctialis Linnaeus. . . . . . . O42 Parkinson's Petrel Procellaria parkinsoniG.R.Gray . . . . . 647 Pediunker Adamastor cinereus(Gmelin) . . . . . . . . 649 xii CONTENTS Pink-footed Shearwater Puffinus creatopus Coues Flesh-footed Shearwater Puffinus carneipes Gould Greater Shearwater Puffinus gravis (O'Reilly) Gray-backed Shearwater Puffinus bulleri Salvin Sooty Shearwater Puffinus griseus (Gmelin) Short-tailed Shearwater Puffinus tenuirostris (Temminck) Manx Shearwater Puffinus puffinus puffinus (Briinnich) . The Least Shearwaters . Tristan Shearwater Puffinus assimilis elegans Giglioli and Salvadori . Audubon’s Shearwater Puffinus lherminieri lherminieri Lesson Galapagos Shearwater Puffinus lherminiert subalaris Ridgway Great-winged Petrel Prerodroma macroptera macroptera (A. ey Black-capped Petrel Prerodroma hasitata (Kuhl) . Galapagos Petrel Prerodroma phaeopygia phaeopygia (Salvin) . Juan Fernandez Petrel Prerodroma externa externa (Salvin) Atlantic Petrel Prerodroma incerta (Schlegel) Kerguelen Petrel Prerodroma brevirostris (Lesson) Kermadec Petrel Prerodroma neglecta (Schlegel) South Trinidad Petrel Pterodroma arminjoniana (Gi glioli and Salvadori) Soft-plumaged Petrel Pterodroma mollis mollis (Gould) . The Small Petrels of the Genus Pterodroma Cook's Petrel Pterodroma cookii orientalis Murphy Mas Atierra Petrel Pterodroma cookii defilippiana (Giglioli and Salvadori) Mas Afuera Petrel Prerodroma leucoptera masafuerae Lonnberg White-throated Petrel Pterodroma leucoptera brevipes (Peale) . Blue Petrel Halobaena caerulea (Gmelin) The Storm Petrels, Subfamily Hydrobatinae Least Petrel Halocyptena microsoma Coues . Galapagos Storm Petre] Oceanodroma tethys tethys (Bonaparte) Peruvian Storm Petrel Oceanodroma tethys kelsalli (Lowe) Madeiran Storm Petrel Oceanodroma castro castro (Harcourt) . Bangs’s Storm Petrel Oceanodroma castro bangsi Nichols . Leach's Petrel Oceanodroma leucorhoa leucorhoa (Vieillot) Markham’s Storm Petrel Oceanodroma markhami (Salvin) Hornby’s Petrel Oceanodroma hornbyi (G. R. Gray) Black Petrel Loomelania melania (Bonaparte) . Gray-backed Storm Petrel Garrodia nereis (Gould) Wilson's Petrel Oceanites oceanicus oceanicus (Kuhl) Fuegian Petrel Oceanites oceanicus chilensis Murphy Elliot’s Storm Petrel Oceanites gracilis gracilis (Elliot) . Lowe's Storm Petrel Oceanites gracilis galapagoensis Lowe . White-bellied Storm Petrel Fregetta grallaria grallaria (Vieillot) CONTENTS alr Pace Tristan Storm Petrel Fregetta grallaria (tristanensis ?) Mather £50) oa alia Black-bellied Storm Petrel Fregetta tropica (Gould) ela clyes aL Oe White-faced Storm Petrel Pelagodroma marina marina(Latham). . . 767 Galapagos Frigate Petrel Pelagodroma marina, subspecies . . . . 77/0 Tue Divine Perrets, Family Pelecanoididae . . .*. . T7/i1 Potoyunco Pelecanoides garnotii (Lesson) . a ee is 3) Magellanic Diving Petrel Pelecanoides magellani (Mathews) : OFA, South Georgian Diving Petrel Pelecanoides georgicus Murphy and Harper ees Falkland Diving Petrel Pelecanoides urinatrix berard(Gaimard) . . . 788 Tristan Diving Petrel Pelecanoides urinatrix dacunhae Nicoll. . . . 790 Coppinger’s Diving Petrel Pelecanoides urinatrix coppingeri Mathews . . 791 ORDER PRUBGAINIFORMES: . -AtiiseYis) utd aie 78 Tue Tropic-Birps, Family Phaéthontidae . .. aD South Atlantic Red-billed Tropic-bird Phaéthon aethereus aethereus Tada 797 Caribbean Red-billed Tropic-bird Phaéthon aethereus mesonauta Peters ey) South Atlantic White-tailed Tropic-bird Phaéthon lepturus (ascensionis?) (Mathews) .

© ‘507. 40>" uo? *) 20° 1GAm 209 *,S07 4.640% eC'50% Y 60S: Fic. 1. Zones of calms, trade winds, and westerly ede in the Atlantic Ocean be- tween latitudes 60° N. and 60° S., with comparative graphs showing the annual means of salinity, precipitation, evaporation, density, barometric pressure, and the temperature of air and surface water. After Schokalsky (1917) with the addition of the atmospheric pressure curve from Schott (1912). [43] 44 OCEANIC BIRDS OF SOUTH AMERICA here in July, and the northeast trade in January. The mean annual rainfall at Buenaventura, Colombia, during a seven-year period (1910-1916) was more than seven meters, with a yearly maximum in October. f. Winds and Seasons in Guiana. The wind régime in Guiana results from a seasonal oscillation of the trade winds and of the zone of equatorial calms from one side of the equator to the other. There are thus two dry seasons corresponding, respectively, to the periods of the northeast and southeast trades, separated by two periods of calms and rains. In Brazilian Guiana the dry season lasts from August to February, rainfall beginning when the southeast winds decline. Farther north a double invierno obtains. At Cayenne the main rainy season holds sway from April until July, the southeast trade restoring a dry cycle from July to December. Then the rains commence again, but during February and March there are almost always a few weeks of fine weather due to the reign of the northeast trade wind, which never reaches as far as the Amazon. In Dutch Guiana, and more especially in British Guiana, the four seasons alternate regularly and are approximately equal in length. The rule of the northeast trade wind, however, increases more and more as one progresses northwestward. At Georgetown the verano of the northeast trade (February to April) is more marked than that of the southeast trade (September and October). The heaviest rains come during May, June, and July, and in December and January. g. The Line of Doldrums. As noted above, the trade wind of the northern hemisphere does not advance at any time of year much to southward of French Guiana. But the southeast trade wind, the northerly limit of which recedes as far as Cape Sido Roque during the austral summer, moves rapidly northward during the northern summer, crosses the equator, and sometimes makes itself felt as far as the mouth of the Orinoco. The mean axis of the zone of calms dividing the trade winds lies, therefore, a few degrees north of the equator. This axis, as measured by baro- metric pressure, resulting wind direction, and precipitation, rather than merely by temperature, is the meteorological equator. The discrepancy between its position and that of the geographic equator is observable equally along the coast of the Guianas and offshore on the Pacific side in the vicinity of the Galapagos Islands. It is one of the phenomena having to do with the unequal divisions of land and water between the northern and southern hemispheres. Over the Pacific Ocean and near the continent, for instance, this line, which marks the doldrums, runs in a generally west-northwest east-southeast direc- tion toward the South American coast, which it meets near Panama during the northern summer and at the geographic equator during the northern winter. In summer a rain-bearing southwest monsoon replaces the southeast trade wind in the oceanic bight at the northwestern corner of the continent (that is, the Cocos-Malpelo-Buenaventura-Balboa area). In winter the northeast trades blow as northerly winds over the Gulf of Panama and almost down to the equator. METEOROLOGY 45 The southeast trade wind blows briskly and regularly along the Atlantic coast as far as the mouth of the Amazon from June to January. During the remainder of the year this part of the coast is prevailingly calm because the northeast trade wind never reaches it, even though it is largely responsible for rainfall in the interior of the continent. h. The Southern-Summer Rainfall Region. South of the equator, except in the anomalous coastal strip of easternmost Brazil, the meteorological conditions are the reverse of those in the north- equatorial belt. In other words, they are entirely homologous with those of the northern-summer rainfall region, for the bulk of the precipitation coincides with the austral summer. During the winter the high-pressure systems over the oceans to eastward and westward widen until they unite across the con- tinent and fill almost the whole belt between the equator and latitude 40° S. Low-pressure systems in the interior of South America appear to be the prime cause of the inward movement of vast quantities of moisture that provide the heavy summer rainfall of the tropics everywhere except along the arid west coast and in the peculiar region of northeastern Brazil to be described later. Over the equator (and between the zones of the trade winds) there is a moisture- laden upper current which moves westward throughout the year above the Amazon Valley, and is largely responsible for the continual rains in the upper part of the basin. When the sun moves southward, this equatorial air is rein- forced by large quantities of moist trade-wind air from both the northeast and the southeast. When the sun moves north of the equator, the southern edge of the Amazonian low-pressure area shrinks as far as about 10° S., and the northern part of the basin then comes under the influence of the equatorial calms. At the same time the northerly coastal slopes begin to be well supplied with rain by the inflow of moisture-laden air from the Caribbean and the North Atlantic. i. The Amazonian Climatic District. The Amazonian climatic district opens out like the small end of a funnel from the valley of the river toward the coast, adjoining the northern-summer rainfall region near the Guiana border, and running across the mouths of the Amazon and down to a point near Fortaleza in the state of Ceara. The wettest months are January to May (on the coast), but rainfall seldom fails at other seasons. As would be expected, it is heaviest in the vicinity of the estuary when the tropical calms prevail, as between December and April. During the southern winter, when strong easterly winds blow in from the greatly enlarged south Atlantic high-pressure area, the climate tends to be fresh and dry. Ina low-lying district at the mouth of the river, as at Salinas, Para, and for some distance along the coast in both directions, is a strip in which the rainfall is distinctly heavy, approaching a mean of 2500 millimeters per annum. Here the distinction between the two annual seasons is not so much that between wet and dry as between very wet and moderately wet. It would appear from this digest that South American summer rainfall bears 46 OCEANIC BIRDS OF SOUTH AMERICA a close resemblance to the monsoon type. Since the continent as a whole, how- ever, is warmer during every season than the surface of the surrounding oceans, there is no such great winter outflow of high-pressure atmosphere as proceeds from the interior of Asia. The only notable complete monsoon system in South America is the small-scale one in northeastern Brazil. j. The Chilean Mediterranean. Two coastal exceptions to the widespread summer rainfall régime of the continent are found, namely, one in this same semi-arid, monsoon region of Brazil, and another in south-central Chile. Both of these widely separated territories receive most of their rain during the austral winter. The latter of the two is the Chilean Mediterranean region, the limits of which are about 27° and 36° south latitude (approximately from Copiapé southward to Concepcion). As in similar regions elsewhere on western con- tinental coasts, this is characterized by winter rainfall, the summer months, especially to northward, being quite dry. Copiapdo has a mean annual precipi- tation of only about 16 millimeters, all of which falls between May and August. We have to go southward almost to Talcaguano before reaching a part of the coast that receives some rain during every month of the year, with the peak (circa 120 millimeters) in June and the annual average approaching 500 milli- meters. The same rule holds offshore, as at the Juan Fernandez Islands, and also southward, with rapidly augmenting total precipitation, as far as the belt of permanent westerly winds. The weather of the Chilean Mediterranean is alternately that of the wester- lies and that of the trades. In the austral winter, when the zone of westerlies has spread equatorward, brisk breezes and moderate rainfall prevail. In summer, when the southeast trade-wind belt shifts poleward, a period of mild, dry, and nearly continuous fair weather sets in. k. The Brazilian Semi-Arid Monsoon District. The second important region of winter rainfall, referred to above, comprises a short stretch along the northeastern seaboard of Brazil, especially in the states of Ceara and Rio Grande do Norte. Here a peculiar semi-arid area extends inland in a southwesterly direction between Cape Sao Roque and the well- watered Amazonian region. It suffers from occasional floods and recurring droughts, and because of its climatic trials, is whimsically known to the Brazil- ians as the ‘‘Job of the North.’’ True monsoon conditions in miniature prevail here. The northeasterly winds that rule in summer are probably the southeast trade wind combined with a summer monsoon, and warped by it from the nor- mal course as soon as the barometric pressure is sufficiently lowered in this part of the continent. The southerly and southwesterly winds that replace it in winter represent the reverse monsoon, which begins when the pressure over the continent rises. The center of least rainfall is found within a narrow strip abutting on the sea between the Assu and Apody Rivers in Rio Grande do Norte. During the METEOROLOGY 47 austral summer, when most of the surrounding territory receives heavy rains, this section is dry. Most of its rainfall comes during the southern autumn and winter, when the remainder of tropical South America, south of the Amazon, has a relatively scanty supply. The explanation of the comparative aridity of this corner of the continent seems to be that the local interior uplands act as a wedge between the sea winds from the northeast and those from the south- east. These winds therefore carry the bulk of their moisture beyond the region, uniting to deposit it in the low-lying central parts of the continent. At the island of Fernando Noronha, well off the coast, and at Natal, Recife, and other points near the easternmost projection of the continent, rainfall is more copious but its annual distribution is similar. The rainiest month is April at Fernando Noronha, June at Natal and Recife. Throughout the whole general region the driest part of the year comes between September and December, when the northern limit of the southeast trade-wind belt is retracting. l. Rainfall on the Brazilian Coast North of Rio. The interior summer rainfall region, referred to above, comprises an enor- mous area in the central and eastern parts of South America, and reaches the sea along the stretch of coast between southern Pernambuco and Rio de Janeiro. The moisture-laden Atlantic air is attracted toward the interior by the great low-pressure system established during the southern summer, and to this west- ward movement the Brazilian plateau appears to offer slight resistance, despite the fact that its highest points lie along the eastern edge. m. Anticyclonic Coasts and Transcontinental Lows. To southward of the trade-wind zone, circulation of the air is regulated by the high barometric constants of the south Atlantic and Pacific. The first determines the south and southeast winds on the northern coast of Chile; the second the northeast winds which run down the east coast as far as the Rio de la Plata. On either side of the continent these winds are stronger and more regular in summer. That is the time when the land is warmed, when the pres- sure is diminished, when the gradient between the high pressure of the oceanic atmosphere and the lighter continental atmosphere is greatest. In the southern winter the pressure increases on the continent until it exceeds that over the surrounding oceans, and the winds become irregular. To southward of 30° S., the normal winds are interrupted by the passage from west to east of cyclonic depressions such as those that cross the North Atlantic. The cold waters and high pressures of the Pacific coast seem to interpose a barrier to them, and they never turn northward west of the Andes. It is for such reasons that storms are unknown along the Peruvian coast. Instead, the lows go to the southward around Cape Horn, or cross the continent directly, reaching the Atlantic between northern Argentina and southern Brazil. Then they usually turn southeast, to the southward of the South Atlantic high-pressure field. How- ever, in winter (May to October), and especially during July and August, at the time when the northeast winds are feeble along the Brazilian coast, these 48 OCEANIC BIRDS OF SOUTH AMERICA depressions frequently run northward as far as the latitude of Rio de Janeiro. They determine the shifting winds (north winds followed by south winds), with sharp variations in temperature. The cold and often violent winds called pamperos follow their passage. Ward (1903, 359) has written: Cyclonic storms are . . . very common in the latitudes occupied by the prevailing westerly winds. The passage of such cyclonic storms across Argentina causes marked changes in wind, temperature, and weather. The warm, damp northerly wind (worte) in front of these depressions is followed by the cool, dry pampero from the southwest, on the rear. The pampero corresponds in general to our own [North American] northwest wind, which is the rear indraft into a cyclonic centre. It is usually cool, dry, and bracing. The exclamation attributed to the first Spanish arrivals in this region, ‘‘Que buenos aires son estos!’’ must have referred to the conditions which prevail with a southerly wind, and not to those which the norte brings. The name Buenos Aires perpetuates, as is often the case with geographical names, a climatic feature of the region in which the city lies. n. The Zone of Westerlies. Beyond 40° of south latitude, or thereabouts, the barometric pressures reduce rapidly toward the polar continent. This is the belt of great west winds—the “Roaring Forties, Furious Fifties, and Shrieking Sixties.’’ Throughout this zone, however, the most important climatic truth is the inconstancy of the weather. We speak as a matter of course of the ‘‘westerlies’’ of these latitudes, but in reality the winds are shifting all the time. Their strength and direction depend much on the season, and at no time are they to be compared in steadiness with such a wind as the southeast trade. It has been shown that in the South Atlantic the maximum line of the westerly winds, determined by multiplying the average velocity of such winds by their frequency, is found between the latitudes of 45° and 50° S. Here the average force is 5.38 of the Beaufort scale and here, too, the percentage of easterly winds is least (Deacon, 1933, 171). In the Pacific the belt of prevailing westerly winds seems to extend farther southward than in the Atlantic. The general course, however, is around the world, and the gusts and gales, interrupted now and then by the passage of lows, always begin again with renewed intensity. Southern Chile, together with the whole of Tierra del Fuego, is under the influence of the westerlies throughout the year. The winter precipitation maximum in this part of the continent is explained by the effect of the Pacific northerly winds which serve to strengthen the westerlies at this season. Wherever there is direct exposure to these winds, rainfall or snowfall is heavy at all seasons, and only in the eastern, low-lying part of Tierra del Fuego does it drop below a high figure on the coast. Altitude is the only requirement necessary to condense still more water from the humid winds of these latitudes, as may be demonstrated by comparing the mean yearly precipitation of Ushuaia, which is 34 meters above sea level, with that of Staten Island, directly to eastward and to leeward, the snow- covered heights of which rear a thousand meters, more or less, above the Atlantic. The respective figures for rain and snow, reduced to millimeters of water, are 479 for Ushuaia and 1701 for Staten Island. The latter has, in fact, even heavier precipitation than that shown by the recorded tables for the island of South Georgia. METEOROLOGY 49 A somewhat similar relationship is that between the Falklands and South Georgia. At both these localities the winds blow chiefly from between south- west and northwest, and gales are frequent. At Stanley the average annual precipitation amounts to 640 millimeters. In the sheltered district of Cumber- land Bay, South Georgia, it is 1400 millimeters; on the high land of the interior, and on the exposed southwesterly coast it must be far greater. Snowfall at South Georgia is heaviest in autumn (May), the period of the year when the barometric pressure is lowest and the storm-tracks are farthest south. 0. The Diagonal Desert Strip. There remains for consideration the Peruvian-Chilean-Patagonian desert region, which is of commanding geographic importance on much of the west coast of South America. This extends along the Pacific from just north of the Gulf of Guayaquil southward beyond Copiapé (27° S.). Then it crosses the Andes, reaching the Atlantic near the mouth of the Rio Colorado (40° S.), and continues down the coast of Patagonia to the Strait of Magellan, beyond which ‘tthe westerly rain-bearing winds make themselves felt right across Tierra del Fuego. The fundamental cause of the aridity in this zone is the high wall of the Andes, which completely cuts off Atlantic moisture in the trade-wind belt, while it is equally effective in intercepting moisture from the Pacific in the westerly belt. Along the arid west coast, rainfall occurs only at long intervals of years, if at all, but mists (garda) resulting from the condensation of moisture at sea are common during the summer months at the seashore and at moderate elevations inland. The cool surface of the Humboldt Current causes the sea winds to condense their water vapor as effectively as rising land. It should be noted here that the coastwise, more or less onshore, winds that blow from a southerly quarter along the reaches of Peru and Chile which are bathed by the Humboldt Current, represent only a deflected moiety of the southeast trade wind. Western continental coasts, especially such as rise steeply to great heights, always have a tendency to pull the trades inshore in this manner. Less than a day's sail to westward from the Peruvian coast, one finds the southeast trade wind blowing steadily on its wonted course. From Copiapé southward the continental temperatures are low enough to make occasional rains possible, particularly in winter when radiation cools the land more quickly than the sea. South of Valparaiso westerly winds become more pronounced and the inshore northward-flowing current ceases to be recognizable as an intercepting factor. From Valdivia southward, the whole Pacific face of Chile has rainfall of over 2500 millimeters per annum, and a nar- row belt extending along the western slope of the Andes receives more than twice that much. In the south Chilean region the climate, inhospitable enough on the mainland, becomes even worse in the archipelago. The precipitation is excessive, the winds fierce and cold, and the sky almost invariably overcast. General aridity, on the other hand, dominates the Patagonian region to leeward of the mountain chain. In response to the increasing force of the 50 OCEANIC BIRDS OF SOUTH AMERICA westerlies toward the south, together with the lessening altitude of the Andes in the same direction, the isohyets tend to turn away from the mountains toward the Atlantic as they approach the Strait of Magellan. The lowlands of Tierra del Fuego are, therefore, relatively humid as well as windy and cold. But northward, the winds blowing across the Patagonian plateau from the Andes are themselves dry and thus tend to absorb any surface water across which they pass. The small rainfall of the Atlantic coast comes mainly during the southern winter and is associated with interruptions of the westerly winds. p. Sea Birds in Relation to Gales and Lee Shores. Long-continued onshore winds of great violence are always possible hazards to pelagic birds, especially to species that do not rest upon terra firma except during their nesting season. In the northern hemisphere the auks and their relatives belong within this category during the large proportion of the year spent to southward of the breeding grounds. In the southern oceans, petrels, fulmars, shearwaters, and terns are the birds most liable to become the victims of such storms. The danger can hardly be called common or widespread, because the birds concerned are as a whole admirably fitted to cope with the worst weather conditions of the region. Nevertheless, in the zone of westerlies and on the exposed windward toasts of South America, Africa, and Australia, vast numbers of sea fowl are from time to time beaten down on lee shores. While we know less of what actually occurs in South America than in the other south- ern continents, one might infer that the destruction in the south of Chile would be worse than anywhere else in the world. Here there is the tremendous stretch of the South Pacific to windward, and the west-wind zone is broader than else- where. To the known strength and persistence of the gales is added the fact that the continent extends southward almost twenty degrees of latitude beyond either Australia or Africa, besides which petrels are extremely abundant in waters between 40° and 60° S. Whitlock (1927, 154; 1931, 263) has recorded the effect of winter storms on oceanic birds along the west coast of Australia, between Bunbury and Perth. His data are referred to in more detail in certain of the biographies that follow. Suffice it to say here that during strong westerly gales he found thousands of petrels and terns driven ashore in a moribund state. On one occasion ten Giant Fulmars and seven of the smaller albatrosses were among hosts of lesser birds, most examples of which had washed ashore alive only to perish later. Whit- lock attributes the weakness of the birds to their inability to obtain sufficient food during prolonged periods of very rough weather. Murphy and Vogt (1933, 325), in discussing the reason why Dovekies or Little Auks (Alle alle) are occasionally blown ashore in great numbers along the Atlantic coast of North America, have pointed out a chain of stimuli and reactions which indicate perhaps more clearly just why petrels and other sea birds sometimes come to grief on lee shores. These authors note first that when beset by such untoward circumstances a bird may lose 50 per cent of its body weight before finally succumbing from exhaustion. They write: SEA BIRDS AND LEE SHORES 51 The physiological processes of birds are extraordinarily rapid, but the katabolic sequence during starvation and exhaustion is the same as among other vertebrates. As the first stage, the glycogen or animal starch in the muscles and the liver is used. This is followed by the combustion of fat throughout the body. Until the glycogen and fat have been practically all utilized, the destruction of muscle fiber is very slight, but thereafter the protein of muscle tissue is drawn upon rapidly for the energy of life and movement. The heart is the last muscular tissue to be diminished. Any starving animal reaches at a certain point a sudden toxic destruction of tissue, which represents the pre-mortal breakdown, and which is irreversible. During such a process of exhaus- tion and starvation the loss in body weight is very great indeed. Since skeleton, skin, and feathers are composed of material that is not combustible, a bird may literally become reduced to “‘skin and bones."’ Possibly as much as 90 per cent of its muscle tissue may be destroyed to supply energy before death ends the process. It is, therefore, not surprising to find that some of the Dovekies had lost half or more of their original weight. The Dovekies, like other pelagic birds, ordinarily waste no energy in fighting the gales of their oceanic range. On the winter feeding grounds they more or less drift with the prevailing wind, slowly while on the water, but becoming part of the more rapidly moving medium as soon as they take the air. Probably their tendency to work to windward prevents undue leeway in normal weather, besides which, winds from new quarters sooner or later bring their compensating force. But doubtless even a very strong and prolonged wind lacks any particular significance for the birds, unless it may eventually cause their planktonic food to descend to deep and inaccessible levels. Ordinarily, the Dovekies would merely move along contentedly and unconsciously with the wind, still within the range of their pasturage, until a leeward coastline set in motion a new series of reactions. The situation is exactly the same as that of the strong-winged Procellariiform species. In the southern oceans many petrels spend their lives in a region of almost perpetual gales. Probably the strongest winds that blow cannot incommode them so long as they have sea room. But when chance and gales bear them down on such a lee shore as the west coast of Australia, for example, they face the unwonted but instinctive experience of combatting the wind, striving to keep offshore. The result is likely to be the same for any species, however powerful on the wing. Rapid combus- tion of tissue is followed by exhaustion, and dying birds are washed ashore by breezes that would have no effect whatever upon them on the high seas (Murphy and Vogt, 1933, 345). q. Birds and Hurricanes. Notwithstanding that practically all of South America lies to southward of the path of Atlantic tropical cyclones, these meteorological disturbances have exercised a profound effect upon the sea birds of the area under consideration. In the first place, the Caribbean-Antillean region is directly to leeward of many late summer cyclones that cross most of the expanse of the northern equatorial Atlantic. This inevitably means that neotropical coasts and waters are more or less peppered each season by stray sea fowl from the eastern Atlantic. This fact alone may go far toward explaining why the Cape Verde Islands, the West Indies, Central America, and the Galapagos Islands share in common one species each of tropic-bird and frigate-bird, and why one booby ranges from the African coast to the Isthmus of Panama. Conversely, we should not expect the American Brown Pelican to reach the Old World (which it does not), because its range in the same belt lies to leeward, with respect both to the trade winds and to the normal cyclonic trends. Secondly, tropical cyclones assist in the transportation into the northern hemisphere of sea birds belonging strictly to the South Atlantic avifauna. A Dy OCEANIC BIRDS OF SOUTH AMERICA striking instance of this is presented below in the case of the South Trinidad Petrel listed under the discussion of Cyclone Number 3 of the selected examples. Since the meteorological equator lies in the northern hemisphere, tropical cyclones never originate in the South Atlantic Ocean. Neither can they have birth within several degrees of the equator, because in such low latitudes the deflecting force of the earth’s rotation cannot be effective in originating a vortex. Nine degrees of north latitude is the southernmost known line of origin of such a movement in the North Atlantic, and most of the cyclones begin in the zone between 10° and 20° N. Since high temperature and high humidity are both favorable, the doldrums possess all the requirements for setting them in motion. The typical course for these storms is first to move westward, then to curve toward the north, and finally to swing in a northeasterly or easterly direction. The movements cannot break through the permanent high-pressure area (anti- cyclone) over the North Atlantic Sargasso region, and hence they skirt to southward of this before recurving, the longitude of the recurve depending upon the position or extent of the high-pressure area. Beginning during the month of August, a large proportion of the North Atlantic cyclones originate far to eastward of the Caribbean region, particularly just south of the Cape Verde Islands, where the doldrums are most in evidence during August and the first half of September. Many of these late storms (September and October) do not finally dissipate until they have reached a high latitude. This is doubtless why most of the tropical sea birds carried into various parts of North America by such movements have been recorded during the latter part of the summer and the early autumn. Cyclonic storms characterized by winds of 100 kilometers per hour or more have an extraordinary effect in transporting sea birds for very great distances. Why this should be so is not easily answered, for, during a large part of their course over the ocean, such disturbances are as a rule of relatively small area. Furthermore, the system of storm winds at sea usually moves forward at a rate considerably less than the average flight-speed of ocean birds, so that the latter should theoretically be able to outstrip the storm center and subsequently to get clear of the peripheral regions of violent rotary winds, dense clouds, and rainfall. Mackaye (in Forbush, 1925, xxii) discusses the matter as follows: These hurricanes, moving north as they do, naturally carry with them any birds flying in their path. It may be assumed theoretically that the influence of such great disturbances on southern sea birds is exerted in the following manner: The wind revolves around the storm center in a direction opposite to the course of the hands of a watch held in a horizontal position. The effect of this revolving gale is felt at a great distance from the center, and the wind velocity is said to reach at times over 100 miles an hour at sea. Probably no bird can long face a gale blowing 60 to 70 miles an hour. As the storm center at sea usually moves first toward the Atlantic coast north of the Caribbean Sea, birds flying over waters that lie south of the center but at some distance from it and under the full influence of the gale would be carried first eastward out to sea, then northward and finally as the storm center moved north along the coast would be driven in from the southeast, east or northeast upon the shore, provided they lived to reach it. Others nearer the storm center might be carried more than once around it before reaching land. Birds migrating at or near the center of the disturbance probably would be caught in the tremendous upward draught there, and might be carried to great heights. Only birds of powerful flight would be likely to BIRDS AND HURRICANES 55 withstand the storm, and small land birdsfiriven to sea by one of these hurricanes probably would never see land again. Sea birds caught near the surface by hurricanes are sometimes crumpled up and driven into the sea. In many cases birds cast ashore during these storms are completely ex- hausted, and in some cases many die of exhaustion after they reach the shore. Such storms probably account for most of the accidental occurrences of sea birds on the coasts of Massachusetts. The above seems to me to be hardly adequate to cover the most remarkable cases of long-distance transportation. From evidence connected with the ‘‘trap- ping”’ of birds within the vortex of a cyclonic storm, and with the occasional conveyance of strong-winged species of sea fowl from the eastern equatorial Atlantic to points far in the interior of North America, I am inclined to believe that an ocean bird might be carried along within the so-called ‘‘eye’’ of a hurricane because, when once entrapped, it would tend constantly to rebound away from the periphery of gales, and thus to retreat toward the quieter center. Let us consider, for instance, the cyclone of late August, 1933, the long course of which is shown in figure 3, and which is Number 3 in the list below. This storm carried a South Atlantic sea bird to central New York State and an eastern Atlantic sea bird to Ontario. Up to August 17, when the movement was still in the eastern Atlantic, the area of gales around the center was probably never more than 125 kilometers in diameter. By August 20, however, the ring of gales had increased its diameter to 600 kilometers, more or less, with the encircling cloud and rain fringe larger in proportion. Now it seems to me alto- gether probable that the birds of the two species referred to were actually caught énside the swirl of this storm. Under such circumstances, they might be carried along without becoming panicky, without experiencing any sense of difficulty, feeding normally, and tending always to turn inward toward the calm of the slow-moving center when they had flown far enough in one direction to come into heavily wind-whipped waters. The system as a whole, by the way, was moving forward during this period at a rate not exceeding 25 kilo- meters per hour. Only when the vortex came into close proximity with the land, as I conceive the situation, would the birds thus held in unconscious durance begin to fight the gales, perhaps to be carried into the higher altitudes of the atmosphere and to be buffeted as helpless waifs for long distances over- land before being cast out centrifugally, subsequently to fall exhausted. It is only through some such process as this that I can comprehend the transportation of Black-capped Petrels from points east of the Caribbean to the Mississippi Valley, or of Madeira and South Trinidad Petrels from the central or eastern north equatorial Atlantic to Ottawa and Ithaca, respectively. Regarding the actual presence and behavior of birds in the vortex of cyclonic storms we have, fortunately, some slight information. Vague accounts of birds seeking refuge upon vessels are likely to appear in the newspapers during every hurricane season. Exact records are rather more difficult to find, but one such has been published by Cradock (1908, 435), and the United States Weather Bureau, through the courtesy of Dr. C. F. Marvin, has supplied me with several others. Admiral Cradock, when in command of a British destroyer, was once obliged to run toward the calm center of a typhoon in the China Sea. When the vessel 54 OCEANIC BIRDS OF SOUTH AMERICA va * bo Fra. 2. Birds, chiefly swallows, seeking refuge on a steamship in the “eye” of a hurricane, Gulf of Mexico, August 27, 1926. Courtesy of the United States Weather Bureau reached a position well inside the whirl of devastating winds, vast numbers of Asiatic kingfishers and other land birds sought refuge upon her. Still nearer the center, where the sky was bright overhead and the atmosphere practically a complete calm—although the waves were toppling about without direction in the utmost chaos—the air was filled with fugitive land birds, fluttering all about the destroyer and everywhere else within the range of vision. On September 3, 1933, the American steamship ‘Golden Horn’ underwent a similar experience in the calm center of a typhoon encountered a short distance to eastward of Shanghai. According to the vessel's log, many different species of birds, including terns as well as land birds, littered the decks, more than 300 being counted upon the bridge alone. Still more graphic, and belonging to our own tropical region, is a record accompanied by photographs and made on board the American steamship “West Quechee’ in the vortex of a hurricane in the Gulf of Mexico on August 27, 1926. In the “‘eye of the storm,”’ at the time the camera was brought into use, the wind was blowing at a rate of only 9 or 10 kilometers per hour. Birds, BIRDS AND HURRICANES 55 which seem to be migrating land birds, chiefly swallows, filled the air about the vessel and were so thickly strewn on deck that they could be scooped up by the armful. The hurricane months in the American tropics are from June to November, with the seasonal peak in August and September. Many of the later distur- bances each season sweep the northeasterly coast of South America. Birds picked up by these storms are in general more likely to be carried northward than toward the interior of South America. Yet, unnumbered waifs must no doubt be blown inland south of the Caribbean, especially during the latter part of the northern summer when so many of the disturbances cross Trinidad and the projecting mainland peninsulas of Paraguana and Goajira. At some future date, when our knowledge of South American sea birds has become less frag- mentary, we shall doubtless acquire plentiful records bearing upon this. Since we have thus far little more than theory to support our conclusions, it will be altogether appropriate to consider specific instances of the extraordinary trans- portation of tropical sea birds into North America by the seasonal cyclonic storms. The question, as noted heretofore, has been treated briefly by Mackaye, but some of the data below will link it up more closely with birds belonging to the South American oceanic area. Before citing several historic cyclonic storms and their associated ornitho- logical phenomena, it may be well to refer to the widely varying rate at which these wind systems proceed along their courses. The speed ranges from less than 150 to an extreme of about 2000 kilometers within 24 hours. These figures represent the total forward movement of the storm, and have nothing whatso- ever to do with the force of the winds blowing around the vortex. In general, the slower the progression of the cyclone along its path, the greater the extent of destruction over any land area, owing to the continuance of hurricane winds at one place for a greater‘length of time. Storm-waves produced by hurricane winds travel, of course, very much faster than the cyclonic system itself, and frequently serve as precursors of what is to follow. The waves and swells of greatest size and length develop in the rear right-hand quadrant of the cyclonic area, and proceed at a rate of 65 kilometers or more per hour in the direction in which the area was moving at the time of their origin. Even the movement of sea birds very often, perhaps usually, out- strips the storm itself, and a surprising number of records in the literature tell of the presence of waifs twenty-four hours or longer in advance of the vortex, or even ahead of appreciable winds. Ober (1880, 179) speaks as follows of the flocks of Sooty Terns which are known as the heralds of hurricanes among certain of the Lesser Antilles: Immediately preceding the hurricanes, there arrive off the Caribbean coast vast numbers of birds called, from their cries, ‘““Twa-oo.'’ They are said to be the harbingers of hurricanes, and only appear during the calms, immediately before a storm. They cover the water in large flocks, and come in from the desolate sandy islands where they breed. They are the sooty tern (the Sterna fuliginosa), but are known to the natives as ‘‘Hurricane-birds.'’ WhenJarrived in Dominica the sea was black with them, but on the morning after the storm they had disappeared, to a bird, as com- pletely as though blown into another sphere. 56 OCEANIC BIRDS OF SOUTH AMERICA : Fic. 3. Courses, so far as determined, of six cyclonic storms responsible for the transportation of pan-tropical sea birds to extralimital regions in North America. Vivian (1904, 55) likewise reports that at Cape Nelson, in British Papua, . a coming “‘blow’’ from the southward was always heralded some hours before by the appearance of a few Frigate-Birds . . . which hovered in the locality while the wind lasted . . . It would be more appropriate to call them ‘‘Prophet-Birds."’ Figure 3 shows the approximate courses of three August and three September or September-October, Atlantic cyclonic storms. Through a comparison of the files of the United States Weather Bureau (Garriott, 1900; Mitchell, 1924) with the ‘A. O. U. Check-List of North American Birds,’ and other sources, these six particular hurricanes have been selected with special reference to their effect in transporting tropical sea birds. (1) August, 1879. A cyclonic storm which barely reached the eastern border of the United States. On August 17, its center was off the Carolina coast. On the morning of the 18th the wind velocity was estimated to be 165 miles per hour (265 k.p.h.) at Cape Lookout, North Carolina, after the anemometer had been carried away. On the 19th the center was off Nova Scotia. During the remainder of the month, Black Skimmers (Rynchops nigra) were recorded on the coast of Maine and in the Bay of Fundy. (2) August, 1893. A very destructive hurricane which, unlike the last, penetrated the eastern seaboard of North America as far as the Appalachian BIRDS AND HURRICANES 57 Mountains. On the coast between Florida and North Carolina more than a thousand human lives are believed to have been lost, mostly through storm- waves. The disturbance has been traced back to its point of origin, near the Cape Verde Islands, about August 18. On the 26th the center passed the Bahamas and next day crossed the Georgia coast. On the afternoon of August 28 the center was northwest of Charlotte, North Carolina, from where it moved in a generally northeasterly direction toward central New York State, travelling 725 kilometers during 12 hours. It subsequently crossed the lower St. Lawrence valley and returned to the Atlantic well to northward of Newfoundland. As a result of this storm thousands of Wilson’s Petrels (Oceanites oceanicus) were washed ashore, dead and dying, along the coast of North Carolina. The 15-kilometer stretch of beach between Beaufort Harbor and Cape Lookout was literally strewn with them (Pearson, 1899, 249). During the last four days of August and the first few of September, moreover, no less than five specimens of the West Indian Black-capped Petrel (Prerodroma hasitata) were taken, respectively, at Blacksburg, Virginia, Oneida and Cayuga Lakes, New York, Pittsfield, Massachusetts, and in Vermont. An example of the Madeira Petrel (Oceanodroma castro) was recovered at Washington, D. C., on August 29, and a Black Skimmer at West Springfield, Massachusetts, about the same date. (3) August, 1933. This disturbance originated to southwestward of the Cape Verde Islands about August 14. It first moved westward for several days and, when northeast of the Lesser Antilles, turned northwestward, passing not far south of Bermuda. It next switched more toward the west, crossed the coast near Cape Hatteras during the night of August 22, and reached Washing- ton, D. C., on the following day. By this time the recurve had begun, and the subsequent movement was north-northeastward toward the mouth of the River St. Lawrence. This storm had a most remarkable effect in the transportation, and inland distribution, of sea birds from both the tropical Atlantic and the north temper- ate Atlantic. The outstanding record is that of the South Trinidad Petrel (Pterodroma arminjoniana) found alive at Caroline Center, near Ithaca, New York, on August 26, 1933. This species had never before been known from North America (Murphy, 1934, 151; Allen, 1934, 134). Hardly less noteworthy is the record of a Madeira Petrel taken alive in the streets of Ottawa, Ontario, on August 28. Furthermore, a Wilson's Petrel (Oceanites oceanicus) was collected on Lake Titus, Malone County, New York, on August 26; 20 Leach’s Petrels (Oceanodroma leucorhoa) were found dead on the shores of Oneida Lake, New York, during early September; and many more examples of the last-named species were scattered over other parts of the Middle Atlantic States. According to the hypothesis advanced at the beginning of this section on birds and hurricanes, it may be assumed that the South Trinidad and Madeira Petrels were transported inside the whirl of the stormfrom somewhere in the eastern 58 OCEANIC BIRDS OF SOUTH AMERICA tropical Atlantic. The same explanation would obviously not fit the records of the Wilson's and Leach’s Petrels, for during August neither of these species would be likely to be present in large numbers in any part of the Atlantic crossed by this particular cyclonic vortex. In this connection it is important to note that the wind at any point in the path of the storm blows from a// directions around the “‘eye.’’ Thus the North Atlantic species scattered over central New York and Pennsylvania by the cyclone of late August, 1933, are just as likely to have been borne inland from the New England coast as from any more southerly part of the Atlantic. The meteorological disturbance as a whole was moving at the time in a northeasterly direction, but the violent winds in its right van were blowing in from the Atlantic to eastward and northeastward. It may, therefore, be essential to distinguish between (1) vortex-borne sea birds, which are transported along the linear path of a cyclonic center; and (2) gale-borne sea birds, which may equally well be carried in a direction form- ing an angle with the course of the vortex. The hurricane of August, 1933, offers the most convincing examples of both categories. (4) September, 1876. This hurricane reached the northerly islands of the Lesser Antilles on September 12, coming from the southeast. On the 13th it crossed Porto Rico and, after skirting the northeasterly edge of the Bahama Bank, turned more toward the north and was due east of Savannah, Georgia, on September 16. Next day the center reached the North Carolina coast, after which it was deflected by a high-pressure area along an unusual inland course, toward the northwest. When it had penetrated well into the continent, how- ever, it hooked about abruptly and swept eastward across Pennsylvania, New York, and New England, and thence out to sea. Between September 16 and 18 the passage of the storm was marked by very heavy easterly gales between Cape Hatteras and Cape Cod. In its wake examples of the Sooty Tern (Sterna fuscata) were recovered at Owasco Lake and Lake Champlain, New York, and at Williamstown and Lawrence, Massachusetts. A Tropic-bird (Phaéthon lepturus cateshyt) was taken at Knowlesville, New York, and a Frigate-bird (Fregata magnificens) at Halifax, Nova Scotia. During the same month, moreover, a Cape Pigeon (Daption capensis) was collected at Casco Bay, Maine, though the presence of this south- ern-hemisphere species cannot definitely be linked up with the hurricane. (5) September, 1878. This disturbance brushed the South American coast and was the most southerly, throughout the early part of its course, of the six selected examples. It originated during August, probably in the eastern part of the Atlantic. On September 1, the center was over Trinidad, from where it turned northwestward across the Caribbean, reaching Haiti on September 4, and running through most of the length of Cuba before the beginning of the recurve carried it slowly northward through Florida and onward up the sea- board of the United States into the maritime provinces of Canada, which it entered on September 13. Heavy precipitation and widespread destruction at- tended this hurricane both in the West Indies and on the continent. Immediately after the passage of the storm, Sooty Terns were reported from BIRDS AND HURRICANES 59 Lake Ronkonkoma (Long Island), and Highland Falls, New York, and from Piscataquis County, Maine. A Brown Booby (Sula leucogaster) was taken on Cape Cod, Massachusetts. (6) September—October, 1898. An example of a cyclonic storm which passed far into the interior of the continent, despite the fact that it crossed the coast well to northward of the Gulf of Mexico. The storm is notable, furthermore, for the complete correlation between its course and the extraordinary records of tropical sea birds in the Mississippi Valley, as noted below. The center passed to northeastward of the Antilles and the Bahamas and struck the continental coast of the United States on October 1, between Jacksonville, Florida, and Savannah, Georgia. It then swept inland far to westward of the Appalachian Mountains and northward through the interior, before looping to eastward across the Great Lakes and passing out to sea south of the St. Lawrence and Newfoundland on October 5. On October 4 and 5, respectively, examples of Prerodroma hasitata were cap- tured at Augusta, Kentucky, and Cincinnati, Ohio. THE HYDROLOGY IN RELATION TO OCEANIC BIRDS Movements within the sea are due to meteorological causes. In a world without breezes, we should doubtless still have a flowing and sorting of ocean waters because of the dynamic effect of unequal heating by the sun. Actually, however, the most obvious of these movements, such as the definite ocean currents, are produced by the secondary agency of wind. Life zones governing the distribution of birds at sea are ultimately determined by physical properties of the surface waters. The well-nigh inexorable control of certain special types of oceanic environments upon birds has not yet been generally realized by either zodlogists or oceanographers. The distributional boundaries and barriers of animals inhabiting land areas, such as mountain walls, deserts, broad rivers, lines of abrupt change in temperature or rainfall, etc., are accepted as a commonplace. Naturalists recognize, moreover, that the ranges of fishes and of innumerable marine invertebrates can be readily cor- related with the temperature and chemical content of sea water. But oceanic birds seem, in the main, to have been regarded somewhat naively as aérial rather than aquatic animals, notwithstanding that their relationships to sea and land, as concerned with feeding and breeding, respectively, are precisely the same as those of the seals among mammals or the sea turtles among reptiles. Members of none of these groups have escaped the necessity of using the land as a cradle, but their true medium, and thesource of their being, is, nevertheless, thesea. In this book we shall have abundant opportunity to note how the majority of oceanic birds are bound as peons to their own specific types of surface water. The bonds are in some cases effective throughout the life of the organism, while in others they apply chiefly to the reproductive period. In a very few instances there appears to be complete freedom from any sort of hydrological control, which means that certain species of birds can thrive equally well in water 60 OCEANIC BIRDS OF SOUTH AMERICA exhibiting the widest possible range of temperature, salinity, and other charac- teristics. But such rare instances only serve to emphasize by contrast the far more familiar circumstances in which the limits of avian ranges suggest that the ocean abounds in invisible walls and hedges, and that trespass, afloat as well as ashore, is in the nature of a zodlogical anomaly. Now in seeking to determine the relationship between ocean water and the composition of bird life in a given district, the first and most useful clue is temperature. Water temperature, rather than air temperature, may be said to govern the distribution of sea birds. The control is rarely a direct one between the warmth or coolness of the water and the sensory system of the bird, although there is evidence of such a simple and immediate relationship in the case of certain penguins confined to the temperate zone, and doubtless among other birds as well. In most instances, however, the control is bound up rather with a long ecologic sequence—with a ladder of phenomena beginning with sunlight and photosynthesis and ending in the nature and quantity of organisms upon which birds may feed. From a biologist’s point of view, temperature tells more than any other measurement we may make about the associated qualities of sea water and the life that it is fitted to sustain. Bigelow (1931, 54) has written: There is as good reason from the biologic side as from the strictly physical for studying the temperature of the sea, because this, more than any other one feature of the water, directly controls the distribution of marine life, animal and plant. Because of the important réle of temperature in governing the rates of animal and plant metabolism, . . . the seasonal changes in the tempeta- ture of the water present special problems to the marine biologist in his studies of important events in the life cycles of animals and plants, such as their breeding periods, the duration of the periods _ of incubation or of larval life, rate of growth, feeding activity at different seasons,.seasonal migra- tions, and many others. The temperature optima and the lethal limits need also to be determined at different stages in development for every species the life history of which is under examination. This question is of practical import in the case of several important food fishes, crustaceans and mollusks: the thermal knowledge that the biologist needs in such cases is, furthermore, of an extremely detailed sort. But, if the quality of the water may tell us something about birds, the birds should also tell us much about water. A Snow Petrel requires water which is cold; a tropic-bird prefers water which is clear, dense, saline, and moderately warm; the Brown Pelican avoids waters that are silty or turbid; the Blue-faced Booby clings to waters inhabited by flying fish which, in turn, are limited to waters of definite temperature and gaseous content. Hardy (1928, 218) has said: In the sea everything is hidden from us. It is only after we haul up our nets, our trawls, our water-sampling apparatus that we can attempt to piece it all together into the geography of the whole and say that here lies a great belt of plankton “‘jungle,’’ there a comparatively barren area, and here again a zone of cold water coming from the Pole. These characters are not fixed like the forests and deserts of the land; but within certain limits are moved by the ocean currents and increased or diminished in size and density by the climatic conditions from time to time. This makes their geography not impossible, but more difficult; at first we can only sketch out roughly the possible limits of these zones. Now one of the objects of this book is to demonstrate that sea birds, which are more easily observed than almost any other animals, offer a ready key to HYDROLOGY IN RELATION TO BIRDS 61 many characteristics of the ocean water and of its hidden life. The correlations are still very imperfectly worked out, and yet a single specimen of an oceanic bird from a remote and little-known island is, in some cases, sufficient to give a broad clue to the characters of the surface waters in the vicinity and the kinds of organisms that inhabit them. Furthermore, as we shall see in connection with studies made at the northern end of the Humboldt Current, birds are sensitively adjusted to, and quick to respond to, periodic changes in the charac- ter of surface waters. In the retreat of certain groups of birds from an area that has formed part of their range, and the invasion of the same region by other species from a different maritime life zone, we can sometimes see a reflection of what is simultaneously occurring in the way of actual replacement of one kind of surface water by another. 1. Tue Nutritionat Basis oF Marine LiFe The organic source of all food in the sea, for creatures of the depths as well as those of the surface layer and of the air above, is the microscopic plant life, comprising mostly the diatoms and brown algae which, obtaining their sus- tenance directly from the nutrient ions in the circulation, build up tissue that becomes the food of small crustaceans, certain fishes, etc., which, in their turn, are devoured by birds and other higher animals. The sun's rays penetrate into the upper layers of the water; oxygen and carbon dioxide are dissolved from the atmosphere and mineral salts are brought in from the land by rivers. These are ideal conditions for plant life; the sea is one great culture medium. Just as the agents favouring life are scattered through the medium, so is life itself; it is scattered as a fine aquatic dust of micro- scopic single-celled plants in untold billions (Hardy, 1928, 211). In the presence of sunlight the microscopic plants assimilate the carbon of carbonic acid and restore oxygen to solution, thus bettering the conditions for animal life. Under optimum circumstances, these plant forms may number tens, or even hundreds, of thousands per liter of ocean water. They exist principally in a stratum within a hundred meters of the surface, though they sometimes penetrate three or four times as far, and their dead remains are uninterruptedly settling into the lightless depths. But so-called “‘vertical’’ circulation, in which as a rule, the angle of ascent is probably very slight, in many parts of the ocean returns masses of deeper water to the surface and thereby prevents food sub- stances in the form of decomposition products from accumulating out of reach of the photosynthetic plant zone. This means that while water denuded of some of its chemical nutrients is carried down, other masses of water that have been enriched during their sojourn below are being restored to the surface. Here, if the food material is not directly available to the algae, it may be utilized by bacteria and worked over into substances which enable the profuse develop- ment of other plant life to go on. Harvey (1928, 165) calls attention to the important fact that, in addition to a supply of phytoplankton suitably spaced in time, a further condition necessary for maximum population is that the energy of plant life passed on to the plank- ton-feeding animals must, before the latter die, be handed in turn to carnivorous 62 OCEANIC BIRDS OF SOUTH AMERICA animals; therefore it cannot be assumed that where vegetable, carbohydrate, and protein food exists there will necessarily be an animal population to eat a fixed proportion of it. In the case of simple death of a phytoplankton cell, writes Allen (1934, 178), it may be supposed that decomposition by bacteria or other saprophytic forms brings derivatives from the body substance to a soluble or suspended condition in the surrounding water, from which some of them may be removed by another cell for its own use. For some particular atom it may be possible for the circuit to be very short, 7. e. diatom—bacterium—sea water—diatom. In most cases it is longer as: diatom—copepod—herring—cod (or bird )—bacteri'um—sea water—diatom. Long or short, there is no room for doubt that photosynthetic organisms occupy the key position in the food exchanges of the ocean. Harvey's diagram of the closed circuit is as follows: ALGAE PHOSPHATES AND NITRATES uaa living on algae oS f a SS GoRPSes AND Carnivorous ANIMALS ————————_> EXCRETA | ANIMALS feeding on detritus BACTERIA ,, ” } Fic. 4 This schema shows that the fertility of an ocean will depend for the most part upon two factors, namely the length of time taken by the corpses of marine organisms and excreta to decay, and the length of time taken by the phosphates and nitrates so formed to come again within the range of algal growth. Where the corpses fall in deep water, well below the light intensity necessary for photosynthesis, and where there is no vertical mixing of the water or deep currents to take the phosphates and nitrates to lesser depths, they are likely to remain lost for many years to the cycle of life owing to the long period which must elapse after the salts are reformed from dead organisms and before they again reach the upper sunlit layers. The process of decay will be slower in the cold bottom water of a deep ocean than in the comparatively warm bottom water of shallow areas. In most places it is the time taken for the re-formed salts to be transported to the upper layers which will predominate, rather than the length of time taken by the dead organisms to decay (Harvey, 1928, 168). In this causal sequence the diatoms are the connecting link between the energy of the sun and the oceanic animal world, including birds. The diatoms surpass in bulk or annual productiveness all other aquatic plants a thousandfold. They are the pasture of the sea; they are highly nutritious and there are perhaps no deleterious species. They are everywhere available, though most so in ocean waters of relatively low temperature, and because of their abundance they repre- sent the sole food supply of certain marine animals, the partial sustenance of many more, and the ultimate source of life for all. Harvey describes the diatoms as single-celled plants, mostly beyond the range of the unaided eye except where they are growing in such abundance as to make visible slimy masses on wet surfaces or in water. Their color is usually a pleasing shade of brown, which shows a peculiar richness when many are massed together. This hue is caused by ‘‘diatomin’’ which covers and hides NUTRITIONAL BASIS OF MARINE LIFE 63 the green chlorophyll necessary for photosynthesis. They are supposed to be allied to the ‘‘Green Algae’’ (Conjugatae) such as the pond-scums and brook- silks. Their outstanding characteristic is a rigid siliceous covering fashioned on the fundamental pattern of a pill box, but much modified in many species. Most botanical texts tell little about them and the encyclopedias are the best references so far as ordinary accessibility is concerned. Enthusiastic diatomists consider the sculptural designs on some of the siliceous coats of sedentary dia- toms as being among the most beautiful things in the world, but these beauties are not so evident in the majority of the plankton forms, which have thinner coverings and more attenuated shapes. Upon these plants feed a host of more or less microscopic animals in which all the large groups in the animal kingdom are represented—if not in adult life then in their younger stages. By far the most important of these animals are the small crustacea; the ocean teems with them, they are the “‘insects’’ of the sea. Pelagic fish, such as the herring, pilchard, and mackerel, and the great whale- bone whales, . . . feed directly upon these plankton animals. From this planktonic world there falls to the sea-bortom a néver-ending rain of dead and dying material which feeds the life of the depths; on the sea-bottom are “‘forests’’ of plant-like animals which, rooted to the ground, stretch out their arms and tentacles umbrella-like to catch this rain of falling food. Upon these, again, feed the creeping animals and the bottom-living fishes (Hardy, 1928, 211). Now low temperature and other characteristics such as low salinity, which is usually associated with low temperature in the southern oceans, make for an abundance of life far in excess of that found in warm sea water. The food sub- stances of all forms of life in the ocean comprise carbonic acid, nitrites and nitrates of calcium and magnesium, etc., phosphates, silica and salts containing a few other elements. These all exist in very small quantities, at most in the proportion of a few parts per million of water. The vast quantity of living substance in the ocean is therefore built up from materials present in exceedingly dilute solution, and the solution is dilute just because organisms are incessantly using up the materials. But sea waters of low temperature are favorable to a high gaseous content and are, moreover, richer in the mineral nitrogenous com- pounds (ammonia, nitrites, and nitrates) than are temperate or tropical waters. The waters of the Antarctic, for example, contain on the average 0.5 per million of nitrogen in the above forms, as compared with an average content of 0.15 per million in the North Atlantic, and 0.10 in equatorial seas. The plankton, and especially the phytoplankton, is therefore far more abundant in polar than in, warm oceans, and especially so in shallow coastal waters of relatively low salinity. Owing to the angle of incidence of a low sun, and the dense screen of plankton, the photic zone is thin. Growth is mostly restricted to the upper hundred meters or so, with much reduced reproduction in the lower strata. The microscopic plant forms, obtaining their food substances directly from mineral sources, combine the simple nitrogenous salts with carbohydrate result- ing from the synthesis of water and carbonic acid under the action of sunlight, and produce proteids. This protophytic type of growth is the basis of the existence of all animal organisms, from tiny copepods to birds and whales. But the abundance of nutritive substance in cold sea water, which has been 64 OCEANIC BIRDS OF SOUTH AMERICA so greatly stressed by oceanographers, does not alone account for the abundance of life. An additional reason lies in the vastly larger number of co-existing generations. Loeb’s (1908, 411) illuminating experiments upon larval sea urchins demonstrate that the temperature-coefficient of duration of life differs enormously from the temperature-coefficient of development. From this he con- cludes that the chemical processes which determine development are altogether different from those which cause old age and natural death. According to the formula deduced by Loeb, if the longevity of an echinoderm larva at T° C. is equal to D., its length of life at a temperature of (T-n) degrees will equal 2" D. In other words, a lower- ing of the temperature n° C. multi- plies the length of life by 2". When the temperature-coefficient of longevity at 10° C. equals 1000, Loeb found the co- efficient of development to be only 2.8. With regard to the extraordinarily rich plant and animal life in the surface waters of the cooler seas, he applied the principle as follows: Within the range of the experiments, reduction in temperature of 10° C. increases lon- gevity a thousandfold; reduction of 20° ; C. increases longevity a millionfold; (iw Taousanos) but the corresponding periods of de- Fic. 5. Mean quantity, in thousands of velopment are multiplied by only organisms per liter, of the collective plankton about three and nine, respectively. ne wp 3. mater of South Acanie From chis it follows that at 0° C the equator and latitude 60° S. many more successive generations of After Hentschel (1933) the same species must exist contem- poraneously than at 10° or 20° C. In tropical seas the predominating groups of marine invertebrates are those which secrete large quantities of calcium carbonate, comprising such forms as corals, macrura, brachyura, anomura, lamellibranchs, gasteropods, etc. In the antarctic seas these are largely replaced by organisms containing little lime, among which may be mentioned tunicates, hydroids, holothurians, annelids, amphipods, and schizopods. Among the last, the Euphausians or opossum- shrimps, which have a non-calcareous carapace, exist in inconceivable myriads and furnish food for vast numbers of vertebrates. The secretion of calcium carbonate is determined mainly by temperature. We see an effect of the cold environment in the feeble development of large molluscan shells or limy skele- tons either in antarctic waters or in the deep sea. Even the shelled pteropods of warm oceans tend to be replaced to southward by naked forms. THE ZONES OF SURFACE WATER 65 2. Tue Zonzs or SurRFACE WATER In the southern oceans, on either side of South America and to southward, there are several kinds of surface water, each zone having its typical amplitude of temperature and salinity as well as other more or less distinctive conditions for the support of plant and animal life. These areas have clearly marked geo- graphic limits and are, naturally, closely related to the general climatic stamp of the respective belts in which they lie. Succeeding one another as they do from polar to equatorial latitudes, we should expect to find a considerable area of water which might be called temperate or Mediterranean but, as a matter of fact, the only region possibly deserving such a designation lies along a section of the Pacific coast. On the eastern side of the continent the diverse tempera- tures of opposing currents (as is true also in the western North Atlantic), and the relatively vast geographic influence of the polar regions upon a “‘water hemisphere,”’ bring sub-tropical and sub-antarctic conditions into such close contact that a temperate oceanic region, in both the hydrological and biological meanings of the term, is practically eliminated. Through the extensive researches of the recent ‘Meteor’ and ‘Discovery’ expeditions, zonal divisions in the South Atlantic have become more precisely known than are those of the South Pacific. Deacon (1933, 171; 1934, 129) has reported upon the four Atlantic regions, called the Antarctic, Sub-Antarctic, Sub-Tropical, and Tropical; Hart (1934, 1) has studied the relation of the hydro- logical circumstances to the associated animal life in the more southerly of these areas. With the conditions in the deeper layers of the ocean, we need not concern ourselves, but in so far as the findings refer to surface waters they bear a highly illuminating relationship to the distribution of South Atlantic sea birds. We may assume, moreover, that substantially similar zonal conditions, more or less modified as to outline and area, hold for the South Pacific. The ‘Meteor’ crossed the whole breadth of the South Atlantic no less than ten times during her epoch-making cruise. If the illuminating charts and graphs published by Hentschel (1933) emphasize any one general point more than another it is the essential harmony in the distribution of primary foodstuffs, microplankton, and all higher organisms, including birds, as determined by the temperature and movements of ocean water. Thus this author's maps showing the relationships of phosphate to plankton abundance, the distribution of plankton in general, the color of sea water in relation to contained life, the order of importance of the several distinct types of phytoplankton pasturage, the abundance of Metazoa of all classes, etc., substantially agree with one another and with the chart indicating the relative numbers of sea birds in all parts of the Atlantic between latitudes 20° N. and about 65° S. (Fig. 53, p. 117). From an ornithologist’s point of view, these data supplement and extend the North Atlantic observations of Jespersen (1930, 14), who found, during lengthy periods of work on the ‘Dana,’ that the regions yielding the least quan- tity of plankton in the surface waters were always those inhabited or invaded by the fewest oceanic birds. While such a conclusion may sound, a4 posteriori, 66 OCEANIC BIRDS OF SOUTH AMERICA like ‘‘an elaboration of the obvious,"’ the ecology of the matter is highly com- plex and deserving of much further research. Layers of ocean water of markedly different temperatures and salinities pre- sent to each other a surface of discontinuity, and behave more or less as though they were separated by a solid wall. Neither heat nor the substances dissolved in the water are readily exchanged between the two. Such discontinuities exist in the open ocean and form, in fact, the physical basis of the zones we are about to discuss. Deacon has shown, for example, that north of latitude 28° S., in about longitude 30° W. of the South Atlantic, there is a very warm layer of tropical water at the surface, separated by a well-marked discontinuity from the underlying water. The surface layer is inhabited by certain organisms not characteristic of the surrounding sub-tropical waters, and is almost depleted of phosphate and nitrate, which can be renewed through vertical mixing only at an extremely slow rate. Deacon also informs me im Jitteris that there are similar layers, depleted in greater or less degree of nutrient salts, in the eastern South Atlantic as well as in the Pacific and Indian Oceans, and that their southern borders are close to the mean annual isotherm of 23° C. for the surface waters. On the other hand, Atlantic sub-tropical water may be carried as far south as latitude 48° S. by the Brazil Current. Sometimes, indeed, one may come upon areas of sub-tropical water entirely surrounded by sub-antarctic water. Such observations indicate that the convergences may be locally sinuous and that they rarely follow definite latitudinal lines for any great distance. The most northerly position of the Sub-Tropical Convergence in the South Atlantic, for instance, is close to latitude 37° S., on the meridian of Greenwich. Near Bouvet Island, typically antarctic waters extend northward far beyond latitude 50° S., whereas to southeastward of Cape Horn they are pushed southward well beyond Jatitude 60° S. In the central part of the South Pacific the Antarctic Convergence is again bent northward, as though by the pressure of waters flow- ing out of Ross Sea. Deacon shows that bottom contours, even at great depths, also affect the courses of the convergences, while the influence of such features of surface topography as the southern tip of South America, and the islands of the Scotia Arc, is even more obvious in throwing these boundaries out of their conventional relation with the parallels of latitude. In parts of the South Pacific, sharp transitions between sub-antarctic and sub-tropical waters, as marked by abrupt horizontal changes in salinity and temperature, are relatively difficult to find. This is particularly true in the east- ern part of the ocean, where the Humboldt Current and related influences tend to obscure the idea] pattern. Observations made during a voyage of the ‘Car- negie’ suggest that there is a clear Sub-Tropical Convergence in about latitude 31° S., at a point as far from the South American coast as longitude 110° W. Vallaux (1933, 177) states that in crossing the Pacific this convergence undulates between the wide limits of latitudes 28° and 40° S., but that in the eastern part of the ocean, which alone falls within our special field, it lies between 30° and 37° S., in so far as it can be determined at all. Later, when discussing the resident sea fowl of Juan Fernandez, San Felix, Tristan da Cunha ph gaye eorgia, f Neti /Falkland | ZSouth Orknéy Macquarie iy 2 R c ty bases Auckland 1. ~ Stewart I. ae — $s Yi [ Crathamis* y/, Y; Miles >_>. 6 sbo 1000 1500 2000 Kilometers 1000 Fic. 6. Zones of surface water, and their convergences, in south-polar projection. The relative zonal positions of the pan-antarctic islands prove highly significant with relation to their respective avifaunas. For example, South Georgia and Macquarie, which lie on opposite sides of the pole in the Antarctic Zone, are inhabited by the same species of diving petrel (Pele- canoides georgicus). The circumpolar islands in the Sub-Antarctic Zone, from the Falklands to Auck- land, share a different species (Pelecanoides urinatrix). [67] 68 © OCEANIC BIRDS OF SOUTH AMERICA and San Ambrosio Islands, and other parts of the eastern South Pacific, we shall have occasion to see that the zonal problems of that ocean are considerably less simple and diagrammatic than they are in the South Atlantic. a. The Antarctic Zone. Determination of the four zones in the South Atlantic is based principally upon temperature and salinity. The Antarctic Zone of surface water is a well- defined, cold, poorly saline stratum, from 100 to 250 meters in thickness, lying above a warmer and deeper layer of water. Its comparative freshness (less than 34%.) is due to the melting of pack-ice in summer and to precipitation which, between latitudes 50° and 60° S., amounts to more than 1000 millimeters per annum, or at least 700 millimeters more than the total extent of evaporation. This layer of antarctic water tends to flow northward at the surface, for reasons stated below, until it meets with the lighter sub-antarctic water, whereupon it descends but continues northward as a deeper “‘creep.”’ In the Antarctic Zone the surface temperature in summer ranges from about —1.0°C. in the south to 3.5° C. at its northern boundary; in winter from—1.8° to about 0.5° C. In the southern part of the zone the salinity may be as low as 33.00%, during the summer. The surface water is extraordinarily rich in nutri- ent salts, the amount of nitrate + nitrite nitrogen running up to 55 milligrams per cubic meter during October and November. The phosphate content never falls below 50 milligrams per cubic meter, so that in this zone there is nothing corre- sponding to the complete utilization of nutrient ions by the phytoplankton of north temperate latitudes in summer. The oxygen content is of the order 90-95 per cent saturation in winter, with frequent supersaturation in spring. Near South Georgia, in January, surface water has been found to be supersaturated with acontent as great as 110 percent. At the same time large catches of diatoms were obtained, and the water had a high hydrogen ion concentration. The lowest oxygen content is found in late summer, when water temperatures are highest. In the Antarctic Zone the effect of the prevailing westerly winds north of 66° S. is to set up drift currents with surface movements toward the northeast and with a total transport of waters in a northerly direction. The increased speed of this surface movement during the southern summer is to be explained not by stronger winds but by the liberation of great quantities of water from melting ice and a resulting thermohaline circulation. From the eastern side of the West Antarctic chain of islands (Graham Land) a nearly uninterrupted southwesterly gale blows out over the ocean, carrying dry and cold air. The Antarctic Convergence with the next zone to northward is also called the Polar Front, and is an important faunistic boundary, dividing the surface waters of two zones. It is indicated by sudden change in surface temperature and is usually found along a line at which the coldest part of the antarctic water sinks below the level of 200 meters (Mackintosh, 1934, 83). It marks the ex- treme northerly limit of pack-ice, although actually the pack is rarely found quite so far north. In summer the line runs not far north of the parallel of 50°S., THE ZONES OF SURFACE WATER 69 at least in mid-ocean. Just north of this convergence there is a region of pro- nounced vertical mixing in the southern half of the Sub-Antarctic Zone. The influence of the antarctic ice extends, of course, much farther outward than that of north polar ice. The northern limit of floe-ice varies greatly from year to year—in the northern part of Weddell Sea it may amount to 1500 kilo- meters. Nevertheless, the differences between the continental and oceanic types of climate are nowhere shown more consistently than between 55° and 62° S. Tierra del Fuego, for example, does not belong to the Antarctic Zone, and is largely covered with dense forests, while Bouvet Island, in the same latitude, is always blanketed with ice and snow, is practically without vegetation, and is surrounded by pack-ice during much of the year. Deacon points out that from a hydrological point of view South Georgia, the South Sandwich group and all more southerly islands, as well as Bouvet, lie in the Antarctic Zone. Further eastward the Crozets and the neighboring islands lie to northward of the convergence, as do also the Falklands. To northeastward of the Falklands the influence of the southward flow of water from the Brazil Current can be seen in a permanent effect upon the surface isotherms. Here the mean temperature is 6° or 7° C., or at least 5° higher than that of the average surface temperature around South Georgia. In a recent lecture before the members of the Royal Geographical Society Dr. D. Dilwyn John, of the ‘Discovery II’ Expedition, has described with singu- lar vividness the climatic and biotic meaning of the Antarctic Convergence. His words are as follows: We began in the west with a line of stations from the western entrance of the Magellan Straits to the south. It ended when we could go no farther, at the edge of the pack-ice. The remainder of the survey consisted of similar lines of stations in a north-to-south or south-to-north direction. The meridional direction is in this area the best suited for our purpose. It is at right angles to the big water movements of the area. Our purpose is to build up a picture of these movements, to define their limits, and to know their flora and fauna. That is to be done most rapidly for a given current by making section after section across it. The southern limit of each line was the edge of the pack-ice. The northern limit was quite as definite a geographical boundary, although it is in the open sea, and not visible to the eye, and not to be defined in a word. One might well ask, what can this boundary be, far away in the open sea? It might be supposed that sea-water throughout the oceans mixed readily, that there would be something like an even and gradual transition from the water of minus temperature, poor in salts because of melting ice and falling snow, at the Antarctic ice-edge, to the warm water rich in salts in the tropics. It is not so: there are successive zones from north to south separated by sharp boundaries. Antarctic surface water is very cold, and although it is poor in salts it is heavier than the warmer more saline water of the neighbouring zone to the north, the temperate zone of the southern hemisphere. This temperate zone is called the sub-Antarctic. Now Antarctic surface-water moves for the most part towards the east because of the prevailing westerly winds, but it has a northerly movement too. Where it meets sub-Antarctic water the very different densities of the two do not allow of ready mixing, and the heavier Antarctic water sinks sharply below the lighter sub-Antarctic and continues its flow northwards below it. This, the line along which the heavy Antarctic surface-water meets the lighter surface-water of the sub-Antarctic and sinks below it, is the boundary in the surface of the open sea that I spoke of. It is called the Antarctic Convergence. It is a physical boundary very easily and precisely detected with a thermometer by the sharp change in temperature as one passes from one zone to another. It can be detected as easily if not so precisely by a zoologist with a tow-net, because 70 OCEANIC BIRDS OF SOUTH AMERICA ance 2 6070S OOOO Fic. 7. Approximate positions of the mean summer isotherms of surface temperature between South America and the Antarctic Archipelago. From Mackintosh (1934). each of the two waters has a distinctive fauna of floating animal life. The zoologist need know only the species of prawns of the genus Exphausia to which Euphausia superba, whale-food, belongs. They are so numerous in the surface that his net will always catch some. If, in the neighbourhood of the convergence, he takes Euphausia vallentini or Euphausia longirostris he is in sub-Antarctic water. He will have crossed the convergence and be in the Antarctic when his net brings back Euphausia frigida and not vallentini nor longirostris. But we, whether sailors or scientists, know and will remember the convergence best in another way: as the line to the north of which we felt one day, at the right season, after months in the Antarctic, genial air again and soft rain like English rain in the spring. I can remember a number of those days vividly. It was like passing at one step from winter into spring. In the southernmost lands in the sub-Antarctic, the islands about Cape Horn, the earth smells as earth should smell and as it never does in the Antarctic. It is no doubt the north-easterly course of the convergence between the longitudes of Cape Horn and South Georgia, so that the former is left far to the north and the latter to the south, that accounts for the vast difference in the climates of two islands which are precisely the same latitude and only 1000 miles apart. The lower slopes of Staten Island are clothed with beech trees with so rich an undergrowth that it is difficult to push through. Darwin compared the richness of the region to that of a tropical forest. South Georgia, the other island, is a true Antarctic land. The snow-line of South Georgia is lower than the tree-line of Tierra del Fuego. We have shown that the Antarctic Convergence is continuous around the Southern Hemisphere. It runs for the most part in the latitude of about 50° S. In the longitude of Cape Horn it lies much THE ZONES OF SURFACE WATER 71 farther to the south than elsewhere. This is because South America has a greater southerly exten- sion than any other land-mass in the hemisphere, and sub-Antarctic water in forcing its way around it pushes the convergence to a high latitude—to over 60° S. In every longitude in the South- ern Hemisphere there is a point at which, going southwards, one finds Antarctic conditions to begin: that is the Antarctic Convergence in that longitude (John, 1934, 383). The southward migration of sea birds in the springtime of the southern hemisphere is, of course, primarily a breeding journey. Yet it must be regarded also.as partly a feeding migration, like that of the whales, for beyond doubt the original occupancy of inhospitable islands and of the antarctic mainland coast by birds was directly related to the rich resources of the surface waters. Within the Antarctic Zone the food of birds, as well as of whales and one or more species of seal, consists mostly of “‘krill,’’ which is the Norwegian whaleman's name for the Euphausian crustaceans or opossum-shrimps. The word, which has been taken over into English scientific literature, has an earlier analogue in “brit,’’ the general English term for right-whale feed in the surface waters of the northern hemisphere. However, since brit refers primarily to copepods and other Entomostraca, krill may well be retained for the Euphausians and related Malacostraca which make up the principal food supplies of surface vertebrates in the Antarctic. * While there are numerous members of the genus Ezphausia, the most abun- dant species, found in practically all stomachs of penguins, petrels, pelagic seals, and whales in the far south, is Euphausia superba. The swarms of this form are the krill par excellence; at times the species makes up the almost exclusive food of penguins (Rustad, 1930, 9). The food of the krill, in turn, consists largely of diatoms which, as Hentschel (1933, pl. IV) shows, make up the predominant microplankton of these waters, with peridineans, Protozoa, and other organisms following. Thus the building up of the bodies of birds and whales is only one state removed from the organic fixation of the radiant energy of the sun by the phytoplankton. (1) BIRDS TYPICAL OF THE ANTARCTIC ZONE In the American quadrant of the far south the following forms are charac- teristic of the Antarctic Zone. PENGUINS PETRELS Aptenodytes forsteri Daption capensis Pygoscelis adeliae Pachyptila desolata Pygoscelis antarctica Priocella antarctica Eudyptes chrysolophus Thalassoica antarctica Pagodroma nivea ALBATROSSES Oceanites oceanicus oceanicus Diomedea exulans exulans Fregetta tropica Phoebetria palpebrata Pelecanoides georgicus Skuas Catharacta skua ‘high’ and 5 “‘low”’ antarctic races) 72 OCEANIC BIRDS OF SOUTH AMERICA These 15 or 16 forms are not of equal individual value as markers of antarctic surface waters. Neither are they all confined exclusively to the Antarctic Zone as breeding birds, while several of them pass very far outside it during their annual migrations. Nevertheless, they all nest chiefly to southward of the Antarctic Convergence, and they comprise an assemblage of quite distinct breeding range from any that follows. A few comments are called for. The three species most closely confined to distinctly polar waters are the Emperor and Adélie Penguins and the Snow Petrel. It is at least possible that a direct response to an extremely low optimum temperature, rather than control of the range solely through special kinds of food, accounts in part for the restric- tion of these species within pack-ice waters. The three species that penetrate farthest south are the Emperor and Adélie Penguins and the Antarctic Skua (C. 5. maccormicki). The last-named, however, migrates during the polar night into the tropics or even into oceans of the northern hemisphere (Peters, 1934, 311). Wilson's Petrel does likewise, while the Cape Pigeon and the Antarctic Fulmar (Priocella) fly close to the equator at least in cool-current regions, as on the west coast of South America. The Macaroni Penguin perhaps nests sparingly at a few islands just north of the Antarctic Convergence, but it is far more characteristic of islands within the Antarctic Zone. The same may be said of the Light-mantled Sooty Alba- tross (Phoebetria). The typical Wandering Albatross is included within this assemblage because, as I shall show later, the form breeding at Tristan da Cunha, at the extreme northern border of the Sub-Antarctic Zone, belongs to a distinct subspecies. A different race of Wilson's Petrel (Oceanites) likewise nests in the Fuegian portion of the Sub-Antarctic Zone. The only breeding whale-bird of South Georgia (Pachyptila desolata) is confined to the Antarctic Zone, at least in New World waters. I believe that Dabbene (1923, 134) was in error when he recorded the definitely sub-antarctic Pachyptila vittata (= forsteri) from South Georgia (cf. also, Peters, 1931, 49). b. The Sub-Antarctic Zone. The Sub-Antarctic Zone of surface water is a much deeper and warmer layer than that of the Antarctic, and is one in which salinity and temperature decrease with depth. It corresponds almost entirely with the belt of westerly winds, and the water in it flows toward the east and northeast, the trend being most northerly in the windiest latitudes, which lie between 45° and 50° S. Conse- quently in the southerly part of the zone, near the convergence with antarctic water, there is a constant tendency toward upwelling in the open sea. Further- more, vertical mixing throughout this zone is promoted by thealmost continually rough weather, making the formation of shallow discontinuity layers impos- sible. Such instability militates against high production of the phytoplankton in the level of optimum illumination. The surface temperatures throughout the Sub-Antarctic Zone i increase, south to north, from about 3° to 11.5°C. in winter and from 5.5° to 14.5° C. in summer. The salinity ranges from about 34%, to 34.5% . The phosphate and nitrate THE ZONES OF SURFACE WATER 73 content of the water is high compared with that found in equivalent northern latitudes, and the presence of from 3.5 to 5.5 milligrams of nitrite nitrogen in the upper hundred meters shows that more nitrate is being generated. The Sub-Antarctic Zone includes Cape Horn, the Falkland Islands, and as much of the Patagonian coast as is influenced by the Falkland Current, which may be regarded as a local extension of sub-antarctic water, and the influence of which occasionally reaches as far north as the mouth of the Rio de la Plata. Gough Island lies within the Sub-Antarctic Zone, but Tristan da Cunha is on the line of convergence with the sub-tropical surface water. On the Pacific side of the continent, the narrow littoral strip of the Humboldt Current, which contrasts sharply with the outlying heated surface water, is characterized to a partial extent by sub-antarctic conditions. The break in meteorological and oceanographic environment is very con- spicuous as a traveller leaves the zone of sub-tropical surface water and enters the sub-antarctic. Quick shifts in the wind are accompanied by changes in temperature more marked than any familiar in the northern hemisphere, while cloudiness, precipitation, etc., vary with equal suddenness. Since many birds and lower oceanic animals are common to both the Antarc- tic and Sub-Antarctic Zones, the term ‘‘Pan-Antarctic’’ has sometimes been applied to all waters south of the mean surface isotherm of about 12° C. (Rustad, 1930, 15). This idea is not without usefulness but, on the other hand, the antarctic and sub-antarctic waters are definitely separable, both in physical characteristics and in floral and faunal correlations. Furthermore, the Sub- Antarctic Zone is itself divisible into southerly and northerly temperature belts which, however, we may disregard for the present. While the Sub-Antarctic Zone is not without Euphausian crustaceans, their numbers rapidly diminish north of the convergence with antarctic surface waters, and their réle as bird-food is filled by other organisms. Conspicuous among the latter are the anomuran crustaceans of the genus Munida, which form a distinct group of ‘‘whale-feed’’ (Matthews, 1932, 481). By the southern whalemen they are called ‘“‘lobster-krill."" The pelagic, post- -larval stage of certain species of Munida, known as the ‘‘Grimothea stage’ because it was originally described under that name as a different organism from the adult, is abundant in the comparatively warm sub-antarctic water around Cape Horn, Tierra del Fuego, the Falkland Islands, and the Patagonian and southerly Chilean coasts. The surface temperatures here range mostly below 10° C., and the creatures occur in enormous shoals, the Grimotheas of Munida gregaria sometimes coloring the surface of the ocean bright red. However, in Humboldt Current waters as far north as the Peruvian coast, where the temperatures are considerably higher, there are also at times vast numbers of Munida. These crustaceans are unknown in the water of the Tristan da Cunha region, which is close to the Sub-Tropical Convergence, and are entirely absent from the much colder waters of South Georgia, Bouvet Island, and the Weddell Sea area. Munida forms a substantial part of the food of many sea birds, including gulls, cormorants, petrels, and penguins. The abundance of the swarms of these 74 OCEANIC BIRDS OF SOUTH AMERICA crustaceans is indicated by the frequency with which they are referred to in the literature of voyages. The ‘Meteor’ records tell of vast “‘clouds’’ of Munida gregaria on the Argentine coastal shelf in latitude 44° S. (Hentschel, 1933, 151). Captain James Cook noted them in the same latitudes more than a century and ahalfago. Still earlier reports are cited by Matthews of the ‘Discovery’ Expedi- tion. Thus Simon de Cordes, Sebald de Wert, and Dirk Gherritz, sailing south- ward along the Atlantic coast of South America on March 12, 1598, wrote,— “having passed Rio de la Plata, the sea appeared as red as blood. The water was full of little red worms, which, when taken up, jumped from the hand like fleas.”’ (1) BIRDS CHARACTERISTIC OF THE SUB-ANTARCTIC ZONE PENGUINS Pterodroma incerta Eudyptes crestatus Pterodroma brevirostris Spheniscus magellanicus Pterodroma mollis mollis Pterodroma cookti defilippiana ALBATROSSES Pelagodroma marina marina Diomedea exulans dabbenena Fregetta grallaria Diomedea epomophora Pelecanoides magellani Diomedea chlororhynchos Pelecanoides urinatrix (2 or Phoebetria fusca more subspecies ) PETRELS CorMoRANTS Halobaena caerulea Phaiacrocorax albiventer Pachyptila forsteri Phalacrocorax magellanicus Pachyptila belchert Phalacrocorax gaimardi Adamastor cinereus Puffinus creatopus SKUAS ae. : Catharacta skua chilensis Puffinus gravis 3 : Catharacta skua antarctica Puffinus griseus Puffinus assimilis elegans GULLs Pterodroma macroptera Leucophaeus scoresbyi A number of other oceanic species might well be added to this list of nearly thirty forms, but I have omitted them because of lack of certainty regarding their zodgeographic affiliations. Furthermore, it would not be inappropriate to include here several birds of the beaches, or common to both salt and fresh water, in southern South America and neighboring insular localities. Among these are a grebe (Colymbus rolland), two or more geese of the genus Chloéphaga, the steamer ducks (Tachyeres), and such shore birds as the two resident oyster- catchers (Haematopus ater and H. leucopodus). The South American Tern (Sterna hirundinacea) is also essentially sub-antarctic, and does not extend southward into the antarctic islands as was formerly supposed. Because of its penetration into cool-current areas near the equator, however, I am reserving discussion of the species for a later section. THE ZONES OF SURFACE WATER 75 A few slightly dubious points qualify the sub-antarctic list as it stands. The distinct race of the Wandering Albatross (dabbenena) breeds, so far as known, only in the Tristan da Cunha-Gough Island district. The Royal Albatross (epomophora) is believed to nest in Tierra del Fuego. Both matters are fully dis- cussed in the biographies. The Dark-mantled Sooty Albatross and the Yellow- nosed Albatross may occupy breeding stations a little to northward of the Sub- Tropical Convergence in the Indian Ocean or elsewhere. Nevertheless, they are typically sub-antarctic birds. Nature is not conventional, and island breed- ing grounds must be used where they are, not where they ought to be. The Galapagos Albatross, for example, nests on an island close to the equator, and yet, as will appear later, it belongs not to an equatorial, but to a definitely temperate oceanic zone. The last aspect also has a bearing on the somewhat puzzling zonal position of a number of petrels belonging to Juan Fernandez, and to the still more northerly twin islets of San Felix and San Ambrosio. The average surface char- acteristics of the Pacific Ocean in the neighborhood of these islands are far less thoroughly known than the corresponding parts of the South Atlantic. A shearwater (Pxffinus creatopus) and several other petrels inhabit one or both of the groups mentioned. Pterodroma cookii defilippiana, for instance, is known to breed at Mas Atierra, Santa Clara, and San Ambrosio. Because of the high- latitude, trans-Pacific relationships of the last-named strongly marked sub- species, I believe that it should be regarded as sub-antarctic rather than sub- tropical. Puffinus creatopus has its closest kin in P. kuhblii, with races in such diverse parts of the world as the Mediterranean-Azorean region and the sub- antarctic Indian Ocean. For the present P. creatopus, which nests as far south as Mocha Island, Chile, may be regarded as a sub-antarctic shearwater, particularly because of the fact that it makes a long trans-equatorial migration, which would be a very unusual trait in a sub-tropical bird. Several other petrels of the Juan Fernandez group, however, I regard as sub- tropical, despite the fact that they nest on a borderline island inhabited by such a definitely sub-antarctic bird as the Magellanic Penguin. The petrels referred to all have affinities with sub-tropical forms in other parts of the world. Their status is discussed below. Certainly no one of the four species of Pterodroma inhabiting Juan Fernandez belongs to the western South American littoral temperate region, to be described hereafter, for the birds properly coming into that category comprise a peculiar and heterogeneous aggregation, all confined to the Humboldt Current. If we had specific information about the food organisms of the various species, the task of zonal allocation would be greatly simplified, but this is a field of inquiry that has barely been touched. Among the sub-antarctic birds listed, three of the shearwaters (Puffinus creatopus, gravis, and griseus), and possibly the skuas, make long, trans-tropical migrations. Other species range widely in the southern oceans, while still others appear to be highly sedentary. 76 OCEANIC BIRDS OF SOUTH AMERICA (2) BIRDS COMMON TO BOTH THE ANTARCTIC AND SUB-ANTARCTIC ZONES PENGUINS Aptenodytes patagonica Pygoscelis papua ALBATROSSES CorMoRANTS Diomedea melanophris Phalacrocorax atriceps Diomedea chrysostoma Gunes ae beens PETRELS Larus dominicanus Macronectes giganteus Sterna vittata Procellaria aequinoctialis SHEATH-BILLS Garrodia nereis Chionis alba The penguins of this pan-antarctic assemblage are remarkable for the breadth of their range. The two species once occupied both South Georgia and the Falkland Islands in great numbers. In former times the King Penguin apparently extended its breeding area still farther southward into West Antarctica, as indicated in the subsequent biographical account. The Gentoo Penguin actually nests a full degree of latitude beyond the southernmost limit of the Ringed Penguin (Pygoscelis antarctica), which is regarded as a species of the Antarctic Zone (Gain, 1914, 52, and chart of distribution). The albatrosses and petrels in the list mostly penetrate during migration well into the cool-current districts of tropical latitudes. The gull even pushes its breeding grounds into low latitudes along both the eastern and western continental coasts, but particularly on the Pacific side. To what extent the exact picture may be qualified by subspeciation is not yet certain. The Kelp Gull of the South Shetlands has been described as a weakly distinct race, Larus dominicanus austrinus (Fleming, 1924, 139). In any event, the range of the resi- dent form of this gull on the west coast of South America is continuous through- out 50 degrees of latitude. In producing such an extraordinary phenomenon, doubtless both the land and the sea exert an influence, and perhaps it is not altogether harmonious to combine littoral with strictly pelagic birds in codify- ing the zonal associations. Larus dominicanus represents an approach toward ““zoneless’’ species of very wide distribution, such as a cormorant, Phalacrocorax olivaceus, which will be discussed later. The question of the two pan-antarctic cormorants also has its unsolved aspects, for it appears that slightly differentiated subspecies of atriceps are restricted to the Antarctic and Sub-Antarctic Zones, respectively. The observations of the ‘Meteor’ naturalists make it clear that we still have a vast deal to learn about the distribution of pan-antarctic and other oceanic birds while they are at sea, far from their nesting ground. Hitherto sea birds have been called either littoral or pelagic, which seemed far enough to carry such a functional classification. But Hentschel strongly suggests that there are degrees of “‘pelagicity.’’ Thus the ‘Meteor’ encountered Macronectes, the Giant THE ZONES OF SURFACE WATER Tr Fulmar, practically everywhere in the southern South Atlantic, while Procel- laria aequinoctialis, although observed hundreds of miles from land, appeared, nevertheless, to be bound to the coasts to a greater extent than its larger rela- tive. During repeated spannings of the ocean, at various seasons, Procellaria was observed northward to 29° S. on the American side, and to 17° S. on the African side where Benguela Current influence is effective. But the species was decidedly more abundant toward the respective continental coasts than near mid-ocean. Not a single example was seen at any time of year between 15° and 25° of west longitude, and very few for a long distance on either side of this strip. Furthermore, the ‘Meteor’ data led Hentschel to the conclusion that there are definite ‘‘gaps’’ along the northern border of the range of albatrosses, into which these birds do not ordinarily enter. He also believes that, although many species of albatrosses are likely to be seen together, most or all species have also special pelagic centers of concentration, which are more or less definite geographic districts, and from which the species concerned tends to exclude its rivals (Hentschel, 1933, 121). For some of these opinions one can imagine sound oceanographic bases, even if they are not yet established. The reason why such a strong-winged wanderer as Procellaria should not pass across the most remote belt of the ocean, after it has already gone so far, is more difficult to fathom. But all parts of the ‘Meteor’ records are deserving of very careful consideration, because of the length of the voyages and the thorough technique of the scientific party. There are sugges- tions here for quite new fields of observation and interpretation. c. The Sub-Tropical Zone. The Sub-Tropical Zone of surface water is much warmer and more saline than the sub-antarctic water. Just to northward of the convergence the temper- ature varies from about 15.5° C. in winter to 18.5° in summer. This increases to an annual average of close to 23° C. at the Tropical Convergence, and like- wise the salinity rises to 36%, toward the north. The nutrient ions show a very striking decrease as compared with sub-antarctic water, and the phos- phate, nitrate, and oxygen content all fall off abruptly toward the equatorial border of this zone. Oxygen, in fact, decreases steadily northward in the surface layers of the Sub-Tropical and Tropical Zones from about 80 per cent of satura- tion in latitude 28° S. to about 40 per cent at the equator. The water is not favorable for diatom growth because of the small amounts of phosphate and nitrate at the surface. Where, however, they are replenished by upwelling, as along the African and Peruvian coasts, the conditions are good. The Tropical Convergence, between tropical and sub-tropical waters, is less well defined than the other convergences. In the South Atlantic it corresponds in summer with the isotherm of 23°C. The boundary passes between St. Helena and Ascension; hydrologically the former is sub-tropical, Ascension, tropical (Deacon, 1933, 171-238). The biotic change between sub-antarctic and sub-tropical waters corresponds with the alteration in physical characteristics. The volume of plankton be- 78 OCEANIC BIRDS OF SOUTH AMERICA comes very much reduced, so that throughout large areas the collective popu- lation of surface organisms may not average higher than ten to each liter of water (Hentschel, 1933, pl. I. Among the microplankton, peridinians and coccolithophores take precedence over the diatoms. Copepods replace the types of pelagic Crustacea that rank first in more southerly zones, but an equivalent richness of individuals is rare and sporadic. The average number of copepods, in both adult and nauplius stages, may range, according to Hentschel, between five and ten to each 4000 cubic centimeters of water. Siphonophores, like the sallee-man and the Portuguese man-o’-war, pelagic gastropods, salps, dolphins (Coryphaena), and flying fishes replace the visible fauna of the colder waters. In general the variety of surface life may be said to increase, while its numbers and total bulk vastly diminish. The count may rise sharply in surface waters adjacent to an island, such as St. Helena, a subject to be reverted to below. The effect upon sea birds of such a local concentration becomes immediately apparent. Between the biota of the Sub-Tropical and Tropical Zones there are few breaks. Rather, the two belts form a warm-water or pan-tropical zodgeographic region extending for a long distance on either side of the equator, as described for the Atlantic by Schott (1926, 281). Therefore a discussion of the sea birds had best be deferred until we have considered the physical conditions of tropical waters. d. The Tropical Zone. In the Tropical Zone the water temperature at the surface ranges between 23° and 29° C., the latter being found just north of the equator. In this line of highest temperature the salinity is at its minimum of 35.5%, because of heavy rainfall and the low rate of evaporation during the season of calms. The light surface water, however, always floats upon water of greater salinity and density, which is observable in cross sections as shallow as 50 meters. The salinity increases to 37%, or more in latitude 15° S. So far as the warmer belt of the South Atlantic is concerned, the western parts tend to become highly saline at the surface because of evaporation due to sun and trade wind. Along the middle part of the Brazilian coast the water may have a salt value up to 37.5%, in spite of much rain at certain seasons. This is considerably higher than the salinity in the eastern part of the ocean at the same latitude. There is no detectable phosphate in the surface waters of the Tropical Zone, and only minute quantities of nitrate nitrogen. Since vertical mixings between the surface and deeper layers are impossible, the probable source of this nitrate is the tropical thunderstorms, flashes of lightning being supposed to cause com- binations of nitrogen and oxygen in the atmosphere. (1) TROPICAL AND SUB-TROPICAL SEA BIRDS The pan-tropical oceanic birds tend to be more local and sedentary than most of the species in the preceding groups. Many must be regarded as more or less land-bound, rather than pelagic, and this applies even to the Procellarii- THE ZONES OF SURFACE WATER 79 formes. Only the tropic-birds and a few of the terns undertake distant travel between seasons of reproduction, and no intertropical species regularly invades the cooler oceanic belts of the northern or southern hemispheres. As an example, Pterodroma arminjoniana, the endemic petrel of South Trinidad Islet, contrasts strongly in this respect with Wilson's Petrel and the Sooty Shearwater, both of which migrate from the cold south to the cold north. In the same way the Fairy Tern (Gygis) contrasts with the Arctic Tern, the latter making annual journeys from the verge of the Arctic to the far south and back. Doubtless the spell of uniformity figures more strongly than the degree of absolute temperature in the control of pan-tropical sea birds. The specific heat of water is so high that in the conflict between air temperatures and water temperatures the ocean wins an overwhelming victory. The marine circum- stances are utterly different from those over rapidly-absorbing and as rapidly- radiating land surfaces. Between the outer borders of the two trade wind belts relative hydrographic and atmospheric monotony is the rule, and over very large parts of the oceans the surface temperature undergoes an annual range of not more than 3° C. (Schott, 1926, pl. XD). To such conditions the life within the sea, and the birds which form an ecologic part of it, have conformed through- out many ages. As hinted above, it is difficult to distinguish between tropical and sub- tropical oceanic organisms, but in the South American region the following five sea birds seem to be typically sub-tropical. PETRELS Pterodroma neglecta Pterodroma arminjoniana Pterodroma externa Pterodroma leucoptera masa- fuerae TERNS Procelsterna albivitta The breeding range of Pterodroma neglecta lies well to southward of the Tropi- cal Convergence all the way across the Pacific, from Lord Howe Island to Juan Fernandez. In the Atlantic its representative species is P. arminjoniana, endemic at South Trinidad and Martin Vas (20° 30’S.). Pterodroma externa is common to Juan Fernandez and the Kermadecs. Prerodroma leucoptera masafuerae is a member of a species with races in tropical and sub-tropical regions of both the southern and the northern hemispheres. The Juan Fernandez form is placed in the fore- going group chiefly because no other sea bird that might be assigned to the Tropical Zone nests at these islands. The Gray Noddy (Procelsterna albivitta) reaches the South American region from southern Polynesia only at the islets of San Felix and San Ambrosio. It is sometimes regarded as a subspecies of the tropical Procelsterna cerulea. I believe, however, that it is rather a well-marked, sub-tropical, representative species, 80 OCEANIC BIRDS OF SOUTH AMERICA with several subspecies of its own. The specific range includes Lord Howe, the Kermadec, Henderson, and Easter Islands. The following 23 birds, several of which are represented by more than one race, belong either to the Tropical Zone or to both tropical and sub-tropical waters of one or both hemispheres. PETRELS Boonies Puffinus lherminieri Sula nebouxii Prerodroma hasitata (including Sula dactylatra P. caribbaea) Sula sula (=*‘ptscator’’) Pterodroma phaeopygia Sula leucogaster leucogaster Oceanodroma castro Sula leucogaster etesiaca Tropic-Birps Eee hing Fregata aquila PEO AER AOES Fregata magnificens PELICANS Fregata minor Pelecanus occidentalis occidentalis Fregata ariel FRIGATE-BIRDs TERNS Thalasseus eurygnatha Sterna fuscata Sterna anaethetus Anous stolidus Anous minutus Gygis alba The list calls for a few comments. Audubon’s Shearwater (Puffinus lherminieri) is an excellent example of a world-wide pan-tropical species, of which two races occur within the South American field. Oceanodroma castro spreads out through an even wider range of latitude, but it appears to be confined to the eastern halves of both the Atlantic and Pacific. Since there is no present connection between the two parts of the range by way of the Indian Ocean, this petrel offers an outstanding instance of discontinuous distribution. In the Atlantic, Oceanodroma castro breeds across the entire belt of the tropics and sub-tropics, occupying one or more islets at each of the following stations: St. Helena, Ascension, Cape Verde Islands, Canary Islands, Madeira, Azores. It is thus quite different in its zoOgeographic affinities from such petrels as Pterodroma mollis and Pelagodroma marina, which have northern-hemisphere races as close to the equator as the Cape Verde Islands (15° N.), but which then skip the tropics and the southern sub-tropics, only to reappear with endemic races at Tristan da Cunha (37° S.). From the latter island Pterodroma mollis, at least, forages toward colder rather than warmer surface regions, and is a typically sub-antarctic bird. In the distribution of the large White Booby (Sula dactylatra) we may note THE ZONES OF SURFACE WATER 81 an exceptionally clear connection with the food of the species. The bird sub- sists principally upon flying fish, and its known breeding stations at the Baha- mas, scattered West Indian localities, the Brazilian Abrolhos, Ascension, the Galapagos, the Ecuadorian coastal island of La Plata, San Felix, San Ambrosio, and elsewhere the world around, all lie close to rich ‘flying fish waters.’’ The ‘Meteor’ chart, showing the relative abundance of these fishes in various parts of the Atlantic, points graphically toward this ecologic correlation (Hentschel, 1933, 114). It is eminently worth noting that the two species of noddies (Anois stolidus and A. minutus) included in the list extend their range southward throughout all the South Atlantic islands to Tristan da Cunha, beyond the Sub-Tropical Convergence. Here, then, we find two thoroughgoing pan-tropical birds shar- ing the same insular station with such equally characteristic sub-antarctic types as the Rockhopper Penguin, the Yellow-nosed Albatross, whale-birds (Pachyp- tila), and skuas. It all goes to show that, since birds must nest on land, and since islands are spaced without reference to any meteorological scheme, the distributional determinants are somewhat elastic in borderline districts. A number of other sea birds, particularly terns, might perhaps be added to the pan-tropical list, but in certain such instances it is difficult to decide whether distribution is not controlled more by continental than by oceanic factors. For this reason even Thalasseus eurygnatha has been included with some hesitation. 3. OcEAN CURRENTS Thus far we have considered the bird life of oceanic zones broadly corre- sponding with the climatic belts of a hemisphere surfaced chiefly with water. Certain irregularities, unconventionalities, and inconsistencies in the system have been hinted at, and the way is now clear to attack such anomalous cir- cumstances in more detail. We shall find that various sorts of minglings, warp- ings, and extensions do, indeed, affect the borders of practically all areas of homogeneous surface waters, resulting in some cases in extreme alteration of the great interparallel plan sketched above. In producing such action, the more stable movements of ocean water, as impelled by winds and deflected by the continental buffers, play the major réle. Thus currents affect every sort of life in at least the upper layers and, in turn, exercise large control over the numbers and distribution of sea birds. Currents can, on the one hand, fence off sterile wastes in the sea while, on the other, they can produce centers of teeming wealth and abundance. In a few instances they are responsible (in so far as, through interaction with living organisms, any environment can be creative) for quite new and distinct aggregations of oceanic bird life, evolved out of material drawn together from widely different climatic sources, and not fitting into any of the marine zones thus far discussed. The Humboldt Current avifauna is an instance of such a result within the South American field. As a preparation for discussing birds and ocean currents together, the geo- physical basis should be briefly outlined. 82 OCEANIC BIRDS OF SOUTH AMERICA Owing to the progressive increase in the eastward linear velocity of the earth's surface from the poles toward the equator, masses (as of air or water) moving in the direction of the equator necessarily fail to maintain an easterly component equal to that of the parallels of latitude at which they are succes- sively arriving. In other words they “‘fall behind”’ or turn toward the right in the northern hemisphere and toward the left in the southern. The converse is true of currents moving away from the equator or in the direction of the poles. These elementary facts explain why the trade winds blow, respectively, from the northeast and the southeast rather than from due north and due south; they explain also why the general circulation of both atmosphere and ocean water tends to be clockwise in the northern hemisphere and counterclockwise in the southern hemisphere. A drift current is produced by the force of the wind upon the surface of the ocean. The direction of water at the surface, however, does not remain in line with that of the wind which caused the movement, but tends to be 45° cum sole from the course of the wind at the moment of impact (cum sole is the direction of the sun's apparent azimuthal motion, 2. e. counterclockwise in the southern hemisphere). For such reasons the southeast trade winds give rise to a surface transport of water in a southwesterly direction. Furthermore, the direction of a surface current of water turns uniformly cum sole with depth, and at the same time its velocity decreases according to a logarithmic spiral until, when the direction of movement has turned through two right angles, it has only one- twenty-third of its surface value. This is the ‘‘depth of frictional influence.”’ The primary results of these circumstances are that in middle latitudes the movement of ocean currents along the east coasts of both North and South America is in the general direction of the poles, (e. g. the Gulf Stream and the Brazil Current), whereas on the west coasts the flow is toward the equator (e. g. the California and Humboldt Currents). It follows that in the Atlantic littoral region tropical or sub-tropical oceanic conditions spread far to north- ward and southward, roughly through 60 degrees of latitude. In the Pacific littoral, on the other hand, coastal waters of high temperature are contracted within a belt barely 30 degrees of latitude in width, owing to the intrusion toward the equator of cool currents from both the north and the south. In fact, the only characteristically tropical waters in contact with any part of the western coast of America lie in a narrow zone between the Gulf of Guayaquil Clatitude 3° S.) and the Gulf of California. Ocean currents profoundly affect much of South America, and in that part of the world exhibit, perhaps, their maximum influence upon the distribution of sea birds. Two of the South American currents act in some degree as exten- sions of the Sub-Antarctic Zone of surface water toward the tropics. These are the relatively short and feeble Falkland Current, which flows northward along the southern part of the east coast, and the great Peruvian or Humboldt Current, which produces highly anomalous biotic conditions along more than half the length of the west coast. Graphic representation of ocean currents can be, at best, little more than OCEAN CURRENTS 83 crudely diagrammatic. As a matter of fact, the entire surface water of every Oceanic area is in constant motion in several planes, such movement being ex- ceedingly variable in certain regions and relatively constant in others. The named ocean currents are merely the vectors of such movement within zones which exhibit a definite permanent or seasonal trend. For a proper understand- ing, the accompanying chart showing the principal currents influencing the character of surface waters around South America should be tempered by con- sideration of Meyer's far more nearly exact approximation of the circulation of South Atlantic surface waters during a single southern-hemisphere summer month. a. The Central Anticyclonic Regions. The ocean currents of the South American maritime area mostly arise from the west-wind drift of sub-antarctic latitudes and from the great counterclock- wise movements around the permanent high-pressure fields of the South Atlantic and the eastern half of the South Pacific. Both of the latter center approximately on the parallel of the tropic of Capricorn. In the Atlantic, for example, be- tween 38° and 30° S., surface water moves eastward under the influence of the wind, carrying Brazil Current water towards Africa. Between 30° and 10° S., water flows westward, as a result of the southeast trade winds, conveying water from the Benguela back toward the Brazil Current. This completes an anti- cyclonic movement extending over the whole width of the ocean. The central parts of these South Atlantic and South Pacific current-rings correspond with the North Atlantic Sargasso Sea. They are relatively lifeless regions, so far as the surface waters go, and are characterized by high tempera- ture, exceptionally high salinity, and by water of an extraordinarily deep blue color and of unsurpassed clarity. The high temperature is naturally favored by the latitude, and by the large percentage of sunny days. Sunshine, in turn, increases evaporation. In parts of the Atlantic area off the coast of Brazil, the salinity sometimes reaches a strength of 38%, while in a corresponding part of the South Pacific, to westward of southern Peru, surface ranges are close to 36.5%. Within these anticyclonic regions there is neither increment of fresh water from melting ice, nor dilution from continental rivers or ocean currents. The latter flow around the ‘’Sargassos”’ rather than into them. While the areas are not entirely without local currents, these have no definitive and constant courses; judging from derelicts and other flotsam, the direction of movement is largely unpredictable, despite the fact that the southeast trade wind blows steadily over a major part of the areas. There is a tendency for the surface waters to accumulate centripetally, and to sink as they become more and more dense, a fact clearly indicated by the high temperatures and salinities of the subsurface layers at least as far down as 400 meters. The exceptional transparency of the water is due in part to causes already stated; all sediments from the coast are, of course, precipitated before they can be carried to such remote reaches of the sea. In the North Atlantic Sargasso, 84 OCEANIC BIRDS OF SOUTH AMERICA £guatorial “Countercurren® aS Fic. 8. Schematic diagram to show the course of the principal current movements named in the text. which may be taken as a type of all such regions, a white two-meter squaré can be seen with the naked eye down to a depth of 66 meters. The paucity of plants and animals is also a contributing factor to the transparency, for such a condition implies very little life either of microscopic plants or of larger organisms. Furthermore, waters rich in plankton usually appear green; those poor in plankton, blue. From an ornithological point of view these central oceanic regions are the deserts of the sea. The important food of pelagic birds is made up of schooling organisms, whether fishes, cephalopods, or crustaceans, and in the anticyclonic centers all such forms of life belong more to the depths than to the surface waters. Therefore abundant bird life is to be sought farther southward, or in the peripheral regions of regular currents and mixing waters. The presence of an oceanic island within one of the permanent high-pressure fields may, to be Die Breite der Plei/schatte bezeichnet dre Bestandigkeit, die PFe/ /befiederung oe mittlere Geschwindigkeit der Stro- mungen. A——A _ sidliche Polartront Subtropische Aon- B B vergenz Vermutliche Hauptbildungs- Statten des poloren Boden - WOSSEPS Gebiete auf steigenden Tie- fenwossers Ww oO ° (« (tty ANN Vi TTT | 7f. UM NA77 77 \ gaan e=_ jem -_ Wige.80°v.Gr70° 60° 50° 40° 30° 20 10° 0° 10° 20°0Lge30°v.Gr. Fic. 9. Flow of the surface waters of the Atlantic in the month of February. From Hentschel (1933), after Meyer. The thickness of the arrow-shafts indicates the relative constancy of movement, while the extent of barbing indicates the relative mean velocity. [85] 86 OCEANIC BIRDS OF SOUTH AMERICA sure, mean a local aggregation of marine birds, sometimes in enormous numbers, but there is a special reason why birds of such colonies can find sufficient food in surface waters close to their nesting ground. The mere shallowing of the ocean on the submarine slopes of islands accounts in large measure for a con- centration of avian food in the only layer in which it can become available, namely the uppermost. This becomes increasingly true if the island happens to lie in a path of current movement, for this compels a constant and plentiful store of food organisms to approach the surface in the vicinity of the birds’ nesting ground. In effect, the food resources of the ocean are improved through temporary reduction in the volume of water beneath a given surface area, and a resultant crowding of the organisms inhabiting it (Murphy, 1924, 231). The soundness of an explanation based upon such an index as the numbers of sea birds is shown by plankton studies of the ‘Meteor’ Expedition. Hentschel (1933, 138-141) concludes that islands without harbors, coves, or appreciable coastal shelves can have little effect upon the quality of life in the sea. As regards quantity, however, he reports greatly increased counts of high-sea plankton in the waters immediately about, or not far from, many of the small Atlantic islands, such as Ascension. Around St. Paul Rocks, which rise from great depths, the ‘Meteor’ party found the largest plankton concentration of an entire cross-section of the equatorial Atlantic. b. The North Equatorial Current. The Equatorial Current skirts the northern margin of the South Atlantic high-pressure area and crosses the ocean in the full sweep of the southeast trades. Because of the closeness of its path to the equator, little deflective force results from the earth's rotation, and the current runs at high speed toward the easternmost projection of the South American continent. In general, a sustained wind produces a surface drift of water amounting to about 1.5 per cent of the wind’s velocity. The speed of the Equatorial Current therefore tends to be least in January, and most during the late summer of the northern hemisphere, the height of the southeast trade wind season. The rate of the Equatorial Current has been a historic matter of concern to - mavigators. Throughout the sailing-ship era it was next to impossible for vessels bound toward the South Atlantic to make headway near the South American coast, or to double Cape Sdo Roque against the current and the south- east trades. The necessary easting was therefore gained in the zone of the north- east trade winds, and the equator was crossed in mid-Atlantic. In November, 1839, Sir James Clark Ross was set far to westward by this current while endeavoring to make St. Paul Rocks. Recently the well-powered German oceanographic ship ‘Meteor’ attempted to describe a circle around these same islets on May 10 but, on account of a northeasterly set of 3.15 kilo- meters per hour, the figure became a pronounced ellipse (Spiess, 1928, 78). The ponderable character of such a flow of surface water becomes still more impres- sive, however, when observed from a fixed base. The following illuminating note was jotted down by Moseley (1879, 68) during the cruise of the ‘Challenger’: OCEAN CURRENTS 87 I never properly realized the strength of an oceanic current until I saw the equatorial current running past St. Paul's Rocks. Ordinarily at sea the current of course does not make itself visible in any way; one merely has its existence brought to one’s notice by finding at mid-day, when the position of the ship is made known, that the ship is 20 miles or so nearer or farther off from port than dead reckoning had led one to suppose she would be, and one is correspondingly elated or depressed. But St. Paul’s Rocks is a small fixed point in the midst of a great ocean current, which is to be seen rushing past the rocks like a mill-race, and a ship’s boat is seen to be baffled in its attempts to pull against the stream. Nearer the South American coast the ‘Meteor’ party measured velocities in the Equatorial Current of 53 kilometers in 24 hours, and off Guiana, in latitude 11° N., of 89 kilometers in 24 hours (Spiess, 1928, 282, 317). At Cape Sao Roque the Equatorial Current divides, and the northern part flows along the northeasterly coast of South America, passing the shores of Brazil, Guiana, and Venezuela, and entering the Caribbean. Throughout this part of its course it is probably the strongest ocean current anywhere within the field of our study. Moreover, as we shall see, its waters, and consequently the life it harbors, are more affected by the outpouring of great rivers than those of any other ocean current in the world. c. The Brazil Current. The Brazil Current flows southward along the eastern face of the continent, varying seasonally in its force and direction. Ordinarily it is strongest in the stretch between the Abrolhos Islets and Cape Frio, where its course is approxi- mately southwestward. Farther northward, as near Bahia, the movement of surface water is sometimes turned in a northerly direction by the trade wind, during the period between May and September. South of the tropic of Capri- corn the Brazil Current becomes so weak that it is of no concern to navigation, though it is still capable of exerting a periodic influence upon the character of the ocean water as far southward as Patagonian latitudes. In general the water of the Brazi] Current is very warm and very blue, which is the equivalent of saying that it is relatively poor in plankton, fish life, etc. The sea bird fauna has in general a West Indian stamp, with the notable absence of the Brown Pelican, a bird which would doubtless thrive along the southerly seacoast of Brazil but which has been excluded from the region by an oceano- graphic barrier to be discussed later. There is also a certain seasonal intrusion of sea birds from the Atlantic Sub-Antarctic Zone. It is interesting to note that even such microscopic organisms as the Fora- minifera of the Brazil Current exhibit the same general type of geographic relationships as the birds. Cushman and Parker (1931, 19) show that along the Brazilian coast, as far south as the harbor of Rio de Janeiro, the foraminiferal fauna is practically identical with that of the West Indies. To southward of Rio, however, a cool-water fauna begins to appear, with numerous species common to the Cape Horn region and to the Humboldt Current zone of the West Coast. Many of the Atlantic warm-water species apparently do not range along the coast much to southward of Sao Paulo, although a few Foraminifera prove common both to Brazilian and Falkland Island waters. 88 OCEANIC BIRDS OF SOUTH AMERICA From such data, as well as from the warming effect of the Brazil Current upon waters not far northeastward of the Falkland Islands, we may conclude that there is a constant meeting, mixing, and biotic struggle between these warm waters and the cool waters of the Falkland Current. The area of contact, however, lies largely in the belt of westerly winds, with a consequent tendency for the surface waters to be carried offshore, and with more or less upwelling along the edge of the continental shelf from cooler layers of intermediate depth. Even when the Brazil Current is exerting its most powerful effect along the coast to southward of latitude 40° S., its climatic influence upon the shores to westward could hardly be a warming one. In fact, the reverse of this might naturally be expected, for, with warmer air over the water off the coast, a stronger flow of cooler air from the pampas would ensue. () THE POVERTY OF TROPICAL BIRD LIFE OFF SOUNDINGS In the Tropical Zone to which we have thus far confined our attention under the heading of Ocean Currents, the relative paucity of bird life on the high sea, when contrasted with conditions in cooler latitudes, is very striking. In general, intertropical sea birds are closely associated with islands, archipelagoes, or con- tinental coast lines. Alexander (1928, 359), after describing the rich variety of tropical sea birds, goes on as follows: The foregoing remarks on the large numbers of species of sea-birds found in the tropics might lead to the supposition that birds were more abundant in tropical seas than in higher latitudes. In reality, however, the opposite is the case. One may travel for days across the tropical oceans and see no birds at all. . . . Even on many tropical coasts the numbers of sea-birds are small. On the other hand many small, desolate, sun-baked islets off the coasts of the continents, and numerous coral islands in mid-ocean, swarm with sea-birds whose cries may be heard for miles across the water and which may appear at a distance like a column of smoke rising from the island. . Since many of the species of tropical sea-birds breed erratically at almost any time of year such islands are hardly ever without some nesting birds. Even where the colony of a particular species nesting on an island has a definite breeding season, another colony of the same species on a neigh- bouring island may breed in a different month, and in exceptional cases birds of the same species may have two distinct breeding seasons in the year on the same island, or the breeding season may differ greatly in different years. Most of the sea-birds which breed in the tropics do not seem to range very far from their breeding grounds. Richards (1909, 5) likewise emphasizes the exceedingly small number of pelagic birds noted by himself and several watchful fellow naval officers during a protracted cruise in the warmer parts of the Pacific. The average on the high sea figured out at approximately one bird for each 200 kilometers of travel. Brooks (1934, 186) more specifically describes the alternation between birdless regions and those of high avian concentration on a voyage between San Fran- cisco and Tahiti. During the first two days, he writes, one sees examples of the North American coastal species, after which there may be five or six days in which faithful watching will yield only an occasional petrel. Within a few degrees of the equator, when the scattered cumulus clouds become thicker and gtayer, a confused swell makes up, and rainsqualls come and go, birds suddenly OCEAN CURRENTS 89 appear. These may include storm petrels, larger kinds of Procellariiformes, and tropic-birds and, when one is not too far away from some island, multitudes of terns may be sighted. As the steamer moves onward, the birds drop away as abruptly as they had appeared. Such alternating phenomena, continues Brooks, are due in the main to the interrelations of currents and countercurrents, along the boundaries of which a swirling and upwelling of water fertilizes the surface. In a cold ocean such as the North Atlantic, the eddies produced by the meeting of such currents as the Gulf Stream and those flowing from the Arctic, cause a similar effect with relation to bird life in certain localities. Instead of the erratic whirls and rainsqualls of the doldrum belts, however, we find rather the production of fogs and slicks. The composition of the bird life, moreover, is somewhat different, but the principle is the same. The ‘Meteor’ investigations show that in the South Atlantic a definite bound- ary between regions rich in pelagic birds and regions poor in pelagic birds crosses the ocean between latitudes 20° and 25° S., from the Benguela Current on the African side to a point south of Cape Frio on the coast of Brazil (Hent- schel, 1933, 116). On the other hand, the ‘Meteor’ chart of the birds system- atically recorded during watch periods throughout the voyages indicates at least two centers or tracks of relative abundance in parts of the ocean within ten degrees of the equator. One of these is in a tongue of water stretching 900 kilometers northeastward from the mouth of the Amazon. The other and more important area lies along a northeast-southwest line corresponding approxi- mately with the shortest distance between Africa and South America. The latter is the part of the ocean which some geographers designate as an “Equatorial Strait’’ connecting the northern and southern moieties of the Atlantic (Vallaux, 1933, 377). Its breadth is but 1530 sea miles or 2840 kilo- meters. Moreover, it is dotted with several islands that support large numbers of nesting oceanic birds, namely, St. Paul Rocks, Fernando Noronha, and the Rocas Reef. Finally, as Hentschel points out, it is also on the line of fairly heavy maritime traffic, which may exert an influence in drawing some of the birds farther than usual from Jand. The avian concentration is thickest between the African coast and about 20° of west longitude, doubtless owing to the effect upon food of a complex mingling of currents in the ocean to westward of Senegal. In the oceanic high-pressure areas discussed above, bird life is, of course, extraordinarily scarce. Much of our knowledge of the southern-hemisphere anticyclonic centers proceeds less from direct observation than from analogy with conditions in the North Atlantic Sargasso. The latter is notorious as a region poor in life (except in the depths). Thirteen species of surface fishes represent the total number observed by Dixon during many years’ observation. The same author also quotes Darwin, who wrote, ‘‘There was more variety of life in and about one leaf of Kelp growing about the shores of Tierra del Fuego than in the whole Sargasso Sea’’ (Dixon, 1925, 441). In bird life, the entire region is singularly poor, as I myself learned during a lengthy voyage through the axis of the area from a point near Bermuda to the Cape Verde Islands. 90 OCEANIC BIRDS OF SOUTH AMERICA RELATIVE FREQUENCY OF BIRDS DURING IO EQUAL OBSERVATION PERIODS 20°*-30°WEST LONGITUDE LATE S52S: Fes.is26 | LAT. 22°S. vuty 1926| LAT. 3° S. JAN. 1927 -w— —~- -w—™~ CN —_~ ALBATROSS CAPE PIGEON STORM PETREL OTHERBIRDS Fic. 10. Relative abundance of sea birds in three latitudinal zones of the South Atlantic. After Spiess (1927). Moreover, ona straight course from north of Madeira to Sa0 Luiz do Maranhdo, Layard, an excellent observer, saw almost no bird life, and his sea captain, who made the same trip regularly, reported that he had never observed conditions otherwise. On another occasion Layard sailed from Para on March 17, en route for Lisbon. Between the mouth of the Amazon and a point within 100 miles of Madeira not a single bird was seen (Layard, 1872, 336; 1873, 331). The work of Jespersen during the ‘Dana’ Expedition and other cruises is still more specific. This author confined his records to sea fowl observed at a mini- mum distance of about 90 kilometers from land. The fewest birds were noted toward the center of the Sargasso Sea, where the average was sometimes as low as one bird per day for periods up to 99 days. In the same area the records reveal the largest number of days during which no birds at all were sighted. Such data are balanced against adequate plankton counts from the surface water, with the expected correlations (Jespersen, 1930, 9). Most groups of sea fowl may be wholly or partly ruled out as being in the Strict sense pelagic within the intertropical zone. Certain migrants from the north or the south, such as petrels, terns, jaegers and possibly skuas, as well as OCEAN CURRENTS 91 phalaropes, regularly cross it along more or less fixed routes, but the truly indigenous pelagic species are reduced to a few petrels, terns, noddies, boobies, and tropic-birds. The last-named are the only members of the Pelecaniformes which are pelagic in the same sense that petrels are pelagic. Cormorants and pelicans, of course, are bound to their insular or continental coasts to such a degree that they seldom lose sight of land. Boobies wander somewhat farther, and yet even they are very rarely seen more than one or two hundred kilometers from shore unless they are crossing a continental bight, such as the Gulf of Guinea. The Brown Booby (Sula leucogaster) appears to be the member of the family most likely to be sighted over the high seas. Finally, the man-o’-war birds, which have such a roaming reputation, also prove, when the data are critically sifted, to cling surprisingly near either their nesting ground or the continental shore line. This is natural, considering that the man-o'-war birds depend for their food partly upon robbing other sea birds; nevertheless, they have a pelagic repute among sailors and travellers which is widespread and hard to eradicate. Now Jespersen’s prolonged observations yielded but one man-o’-war bird as far as 500 kilometers from the nearest West Indian island, and only one other as far as 350 kilometers from land (1930, 26). Hentschel (1933, 122) reports after many crossings of the South Atlantic in the ‘Meteor,’ *’ Fregata habe ich nur in der Nahe von Kiisten und Inseln . . . gesehen.’’ Most of the sup- posedly offshore records in the Pacific have also been made, as a matter of fact, at points close to one or to several islands. It should be remembered that a man- o’-war bird out of sight of land does not necessarily mean that the land is out of sight of the bird. Such facts are easier of comprehension when we stop to consider that no less than four species of Fregata, namely aquila, magnificens, minor, and ariel, and apparently several more subspecies, occupy islands lying in warm oceans within easy flight range of one another and of South America. From this we may infer that while these birds obviously possess the means of surviving, and of estab- lishing new colonies, when they are involuntarily transported long distances, as by storms, their habits are, nevertheless, relatively sedentary. Therefore, when Walt Whitman wrote concerning the man-o’-war bird— At dusk that look’st on Senegal, at morn America the only flight involved was a poetic one. Tropic-birds (particularly Phaéthon lepturus) are, however, frequently ob- served in the remotest parts of the warmer oceans, and are even characteristic of the barren anticyclonic areas. Yet their small absolute numbers, as compared with the multitudes of petrels to be seen in high southern latitudes, are a fair index of the relative availability of food in the surface waters of the respective zones. Most oceanic birds feed upon freely moving organisms, and cuttlefish make up the bulk of the food of tropic-birds. The ‘Dana’ naturalists found no other organisms in the stomachs of eight of them taken far from land. Cephalo- pods, of all oceanic invertebrates, make up the greatest single food resource 92 OCEANIC BIRDS OF SOUTH AMERICA common to widely varying zones. In cold waters the chief combination of importance to birds is probably cephalopods, crustaceans, and small schooling fishes; in warm waters cephalopods and flying fish take high rank. But cephalo- pods and fish are at least ten times as abundant in the surface waters beyond latitude 40° S., and in cold coastal currents, as they are within what we may call collectively the Sargasso regions; and the corresponding ratios with refer- ence to crustaceans are nearer ten thousand to one. d. The Cape Horn Drift. The general eastward drift in the southern Pacific, which flows along a median axis of about 50° south latitude, is impeded by the long poleward ex- tension of the South American continent and is deflected to southward of Cape Horn. The resulting current through Drake Strait, abetted by the Burdwood Bank and other contours of the ocean bottom east of Tierra del Fuego, keeps bergs and floe-ice from reaching waters around the Falkland Islands. Farther eastward in the Atlantic, bergs are familiar in lower latitudes: in mid-ocean, near the Tristan group, as far north as 37° S.; in the Cape region of Africa up to 35° S. They are practically unknown to westward of the Falklands. A few records from the Patagonian coast, off Bahia Blanca, perhaps represent bergs that have drifted inshore after passing well to eastward of the Falklands. After rounding the southern tip of the continent, the Cape Horn Drift divides, one branch looping northward on both sides of the Falkland Islands, while the other and larger passes to eastward on both sides of South Georgia. Hardy, of the ‘Discovery’ Expedition, has explained the peculiar effect of this current in making South Georgia one of the world’s great centers of whale population. His remarks apply equally well to the amazingly abundant bird life of that island which, it will be remembered, belongs to the Antarctic Oceanic Zone, despite its relatively low latitude. Hardy (1928, 220) writes: We found that all the Euphausians were concentrated along the north-east side of South Georgia. Immediately at each side of the island was a zone some 15 miles broad very poor in plankton, and outside this an encircling zone of thick plant plankton, largely diatoms of the species Corethron valdiviae, with outside this again an area of more mixed but less dense plankton. We may guess at the explanation of this. South Georgia is a long narrow island, some 100 miles long by some 15 across; it is placed almost at right angles to the main westerly drift coming up from the Drake Straits. The currents set up round it will be like those set up round any long object forced sideways through a fluid—the water will be forced in a curve round either end to meet in an eddy some distance behind it, leaving am area of “‘dead'’ water immediately against the land. The growth of plant life in the sea is limited . . . by the quantity of available phosphate. This as the summer advances gets used up in the upper layers. Here, where the main ocean current from the west strikes the continental shelf of South Georgia, there will be an upwelling of water rich in phos- phate from the deeper layers on the west side of the island. OCEAN CURRENTS 93 It is here that we get the densest growth of diatoms, which are carried round either end into the area behind theisland. . . . Here in this sheltered water are all the Euphausians, young and old, which feed directly upon the diatoms. It is like a sheltered nursery for them, supplied by food on either side. This theory appears to fit in very well with the results of phosphate analysis obtained by Mr. Clowes, our hydrologist, and is perhaps the explanation why South Georgia, so peculiarly situated, should be one of the richest whale-feeding grounds of the world. The enriching effect of the presence of islands is, of course, reflected by the bird life in the same latitudes. The ‘Meteor’ naturalists found relatively few birds on the open reaches between the South Shetlands and Bouvet Island; at times even the nearly ubiquitous Cape Pigeon failed to appear in the daily census. Supplementing the factors described above by Hardy, Hentschel (1933, 141) also refers to the favorable effect of abundant bird guano upon plankton in the coastal waters of such an island as Bouvet. The influence of the modern antarctic whaling industry upon bird population and bird concentration is dis- cussed elsewhere, in the section relating to petrels and albatrosses. Suffice it to say here that in enclosed waters occupied by whaling shore stations, such as the crater harbor of Deception Island, counts of the zo6flagellates sometimes show more than a million to each liter of water (Hentschel, 1933, 156). Thé Cape Horn Current brings much ice from Bellingshausen Sea, on the western side of West Antarctica, into the South Atlantic and, together with the chilled winds from the same direction, is responsible for the rapid fal] in temperature experienced in passing from west to east in this belt. Thus the annual mean at Ushuaia is 6.2° C., that at Cape Pembroke, the eastern extremity of the Falkland Islands, 5.9° C., and that at South Georgia 2.1° C. (Brooks, 1920, 98). Such atmospheric temperatures well reflect the differences between the Sub-Antarctic and Antarctic Zones of surface water. The west winds and the Cape Horn Drift prevent the additional ice coming out of Weddell Sea from having any direct effect upon the South American region, but the ocean to eastward, toward Bouvet Island and the southern Indian Ocean, is further chilled by it. Many of the results, as we shall see, are closely correlated with the peculiar distribution of certain penguins and other sea birds. Furthermore the conditions are in large part responsible for the fact that the eastern or African side of the South Atlantic is, on the whole, richer in birds and other pelagic organisms than the western side. e. The Falkland Current. The Falkland Current is strongest along the outer edge of the Patagonian coastal shelf, which is the only extensive shallow-water region on the American side of the South Atlantic. Between 40° and 45° S., or thereabouts, it flows northward at the rate of nearly two kilometers per hour, and here its surface temperature is also lowest because of upwelling caused by the prevailing winds from the land, 7. e. the westerlies. On the continental shelf itself the tempera- tures are somewhat higher. The northward extent of the current is variable, as previously indicated, but the area of mixing between its waters, those of the shelf, and those of the drift from the north, are very rich in life and are famous 94 OCEANIC BIRDS OF SOUTH AMERICA for a vast concentration of sea birds at certain times. Here pelagic petrels and littoral gulls, cormorants, and penguins mingle in force. Here, as has already been recounted, Mr. Beck found some of the best collecting grounds of the Brewster-Sanford Expedition. Southeastward of the mouth of the Rio de la Plata, sharp differences in water temperatures are especially notable. Here also cool dry southwesterly winds meet the warm northerly winds that bring thundershowers. Widespread fogs are frequent during the southern winter season, sometimes reaching into the estuary of the great river. There is also much amplitude in the seasonal range of temperature because of the clear skies that generally prevail in the southern summer and the heavily overcast skies of winter. Climatic conditions along the coast are somewhat similar to those of the Newfoundland region in the north, where warm and cold currents likewise come into contact. But Patagonia, as Schott has said, stretches practically from the polar zone to the tropical. f. The Humboldt Current. In crossing the breadth of the South Pacific the Sub-Tropical Convergence undulates between about 28° and 40° of south latitude. In the middle reaches of the ocean the line falls between 30° and 37° S., but midway between Easter Island and Juan Fernandez it loops again northward. Still nearer the continental coast it loses definiteness and significance because of new oceanographic complexities. The heart of the permanent anticyclone lies north of the convergence, be- tween the neighborhood of Easter Island and the continent, its longer axis closely paralleling the tropic of Capricorn as the whole center fluctuates pole- ward or equatorward with the changing seasons. On its southern side is the zone of westerlies, to northward that of the southeast trades, and to eastward the coastwise winds which, near the land, show the usual onshore and offshore diurnal alternations, although always proceeding from a southerly quarter. The oceanic wind system, therefore, tends to blow counterclockwise around the periphery of the high. The region as a whole is prevailingly tran- quil, with transparent, deep blue water and clear horizons; clouds in these latitudes pertain mostly to lofty coasts and the peaks of islands. It is, indeed, the particular part of the South Sea that earned Magellan's appellation of *‘Pacific.”’ The courses of surface movements of the water are, naturally, correlated with the winds. Of the Sub-Antarctic West Wind Drift we need add nothing to what has been said above, but to eastward of the high-pressure center, and particularly in the littoral belt of the greater part of the west coast of South America, are surface flows which determine the entire complexion of the marine avifauna, not to speak of their broader effect upon climate, hydrology, and the life of the land as well as of the sea. Moving northward along the eastern border of the high-pressure nucleus, but still far from the coast, is a current of blue, moderately warm water, which should perhaps be regarded as merely the offshore edge of the cool coastal current but which Kerhallet, in 1853, named the Mentor Current. The discoverer OCEAN CURRENTS 95 computed its velocity in latitude 26° S. as from 18 to 21 sea miles per day, or approximately 1.5 kilometers per hour (Vallaux, 1933, 195). Zorell (1928, 166) has maintained that the Humboldt Current should be regarded as only a relatively cool oceanic stream, holding that the very low temperatures encountered along the tropical coast of Peru, and at times even of Ecuador, are only an incident among many phenomena associated with a broad and deep body of water. This leads me to say that there are doubtless two senses in which the term Humboldt Current may be used. One of these is climatic or oceanographic, and, from this point of view, characteristics of the water hundreds of kilometers from the shore, where the temperature is always high, may be as necessary to an understanding of the current as knowledge of the surface temperatures close to land. The other sense is strictly biotic and zodgeo- graphic, and in the present connection mainly ornithological. Since it is the indubitably cool coastal water that is responsible for all the peculiarities and anomalies of sea bird distribution in the region, it should be made clear that by Humboldt Current I mean specifically the narrow littoral zone of upwelling, together with its northwesterly seaward extension near the equator. The cool coastal current takes its origin not from latitudes bordering the Antarctic, but in the west wind zone, from where it passes into the sphere of the trade winds. These, as previously noted, become deflected to southerly or even southwesterly winds along the steep Andean seaboard. The current is observable along the coast from the vicinity of Mocha Island, Chile (38° 30’ S.) northward to Cape Blanco, Peru (4° 27’S.). From the latter general neighbor- hood it sets west-northwestward, flows on both sides of the Galapagos Islands, and is lost beyond longitude 100° W. in the south equatorial drift. Significant features of the surface waters in the current are first, relatively low temperatures in close proximity to the land, with rising temperatures off- shore along lines usually perpendicular to the trend of the coast; and second, extraordinary uniformity of temperatures throughout the greater part of the length of the current, a uniformity which is little affected either by latitude or season of the year. Both of these facts would strongly suggest that the low temperatures close to shore are due to upwelling from cooler intermediate layers, rather than to northward transportation of sub-antarctic surface waters. The latter would, of course, become gradually warmed during their progress into the tropics, and the Humboldt Current would show appreciably rising temperatures from south to north, which is not in accord with the facts. Up- welling would, in any event, be inevitable in view of the meteorological régime. An accelerated left-hand trend, and continuous vertical circulation, is caused by the steady southerly winds parallel with the coast, which tend to force the surface water offshore at an angle of 45° from their path. The Humboldt Current flows at velocities of from .3 to .6 sea miles per hour, or occasionally more, off Peru, increasing to from 1 to 2 sea miles after turning westward toward the Galapagos. The current is of relatively high salinity and is most strongly marked during the southern-hemisphere winter. The isotherms of surface water are usually parallel with the coast line, but occasion- 96 OCEANIC BIRDS OF SOUTH AMERICA ally they lie east-west because of upwelling from cool substrata. Sharp bends in the continental outline, as at Arica, enhance and complicate these effects so that the northerly coastal waters are frequently of lower temperatures than those to southward. The averages for the Mollendo-Pisco reach, for example, are lower than those off Chile, while during the southern spring and summer the Antofagasta water temperatures are anything up to 3° C. warmer than those at Callao, more than a thousand kilometers to northward. Correlated circumstances are frequently reflected in the atmospheric temper- atures at the Peruvian guano islands. Thus the means for the month of Decem- ber, 1934, at islands succeeding each other from north to south or in a direction away from the equator, were the opposite of what might be expected, to wit: Lobos de Tierra, 19.35°; North Guafiape, 19.68°; Pachacamac, 21.63°; Islay group, 22.5° C. For January, 1935, a corresponding sequence is: Lobos de Tierra, 21.64°; Lobos de Afuera, 21.67°; central Chincha, 22.9°; Islay group, 22.5°; Ifiani, 24.2° C. Part of the effect may be due, of course, to strictly local conditions, but the general trend is significant. The upwelling phenomena responsible for such anomalies begin a little north of Coquimbo and are active as far as Paita. Increasing breadth of the coastal shelf especially favors upwelling, which sufhciently explains the fact of cooler water in the tropical sector. North of Callao, for instance, the con- tinental shelf is relatively wide, whereas along the northerly Chilean coast depths of 200 meters or less are found only within 2 sea miles of shore (Schott, 1931, 161). Coastwise observations during the months of September and November, reported upon by Schott, show that to southward of Paita, Peru, lower surface temperatures are accompanied by lowered salinity, indicating that the source- level of the rising water is deficient in salt. North of Paita, in the equatorial section of the coastal Pacific, the reverse is true. Because of this Schott is entirely in accord with Sverdrup’s (1930, 257) recent findings that upwelling in the Humboldt Current is practically limited to the upper 200 to 300 meters. Average surface temperatures close to shore along the greater part of the coast of Peru range between 14.4° and 17.8° C. (Murphy, 1923, 67), a range some- thing like 10° C. lower than the theoretical value for the latitudes. More extraordinary than the absolute temperature, however, is its general uniformity. As a corollary, the atmospheric temperatures of the coast are maintained at a monotonous level from Mocha Island almost to the Gulf of Guayaquil. Anto- fagasta, Iquique, Arica, Mollendo, Lima, Trujillo, Paita, etc., are all under the effective control of such oceanic influences (Jefferson, 1926, 443). Further- more, the seaward slopes of the mountains, as well as the littoral ocean for more than 150 kilometers from shore along this stretch of coast, are, in the ordinary sense of the word, rainless. Because of these conditions the terrestrial biota is relatively poor, while marine life is both extraordinarily rich and deli- cately adjusted to the environment. How delicate the adjustment is becomes immediately apparent during the infrequent phenomena that temporarily inter- rupt the uniformity. OCEAN CURRENTS 97 Between the westernmost projection of the South American continent and the Galapagos Archipelago, the Humboldt Current normally attains its maxi- mum velocity. In this region, too, the cool, greenish waters intermingle with the much warmer blue waters of the Equatorial Current, the two often forming distinct bands or stripes, separated one from another by rips and foam-lines and accompanied by much turbulence or ‘‘boiling.'’ Such mixing zones often prove very rich in life. The phenomena have been vividly desctibed by Colnett, du Petit Thouars, Darwin, Beebe, and others (Murphy, 1923, 65; Vallaux, 1933, 200). Practically all of the oceanic area bounded by latitudes 5° S. and 3° N., and longitudes 85° and 100° W., is one in which opposing currents fight out a varying but indecisive battle for supremacy. In the zones of confluence and along the boundaries between green and blue water, conditions are favorable for sea birds, as suggested above. Outside these zones, in the tropical expanses of feeble currents and marine stability, few birds are to be looked for. The matter has already been discussed, but it is worth adding that du Petit Thouars, voyaging in the ‘Vénus’ from Callao to Hawaii in 1838, noted especially the great relative scarcity of birds over the tropical high seas (Vallaux, 1933, 204). Birds are, of course, dependent upon marine life that comes very close to the surface. Whales, on the other hand, can pene- trate somewhat deeper layers for their food. This is doubtless the reason why remote blue-water grounds, including even parts of the anticyclonic areas, were profitable hunting centers for the sperm whalers of old, notwithstanding their paucity of bird life. The uniformity produced by upwelling along the continental coast makes the inshore waters of the Humboldt Current an extraordinarily constant en- vironment, as well as a very rich one. No better example of relatively unchang- ing relationship between organism and milieu can be found than that of the diatoms of the current. I collected specimens of these microscopic plants from the surface waters along two-thirds of the length of Peru, and others from ancient pits of guano on the Peruvian islands. Allsamples, regardless of station, season, or year of origin, contained the same characteristic species, which is of special interest because it is not uncommon for the diatom flora of a given local- ity to be strikingly dissimilar at different seasons, and during the same season of successive years (Mann, in Murphy, 1923, 72). An identical principle applies to the metazoan fauna of the Humboldt Current. The population of the surface waters is made up chiefly of a rigidly circum- scribed group of organisms, varying little in constituency but running rather to the development of vast numbers of individuals within a comparatively small number of species. The percentage of endemism is also high. Bigelow’s remarks on the invertebrates are doubly interesting in that they refer both to the abundance of life in the current and to the contrasting conditions in warm- water areas beyond. He writes: I cannot pass over without mention the extraordinary richness of the Humboldt Current in pelagic life of all kinds; a richness which has already been noted in the account of the general oceanographic features of the Eastern Tropical Pacific by its [the expedition’s] leader, who speaks 98 OCEANIC BIRDS OF SOUTH AMERICA of “such masses of Salpae, of Cytaeis, or Cymbulia, or swarms of other pelagic animals as to make athicksoup"’. . . , and of enormous quantities of copepods, schizopods, Doliolum, and Medusae. Nor can J omit to recall the discovery, immediately to the west and southwest of the Current, of an area as barren in all forms of life, bottom as well as pelagic, as the latter is rich. Thus ‘As soon as we ran outside of this [the Humboldt Current] the character of the surface fauna changed; it became less and less abundant as we made our way to Easter Island, the western half of the line from Callao becoming gradually barren.’’ On entering the current again from the barren area the reverse change was equally striking. Paul (1932, 202) and others have emphasized similar phenomena observed during cruises of the ‘Carnegie.’ An impressive picture of the almost unparalleled profusion of living crea- tures, particularly of fishes and birds, is given in literature which I have cited elsewhere, and from which both the above and the following quotations are taken (Murphy, 1923, 80). Coker, for example, has written: In contrast to the barrenness of the coast [of Peru] there is a peculiar wealth of certain forms in the open ocean. The great red seas, formed sometimes, at least, of myriads of microscopic dino- flagellates, are of common occurrence. . . . Sometimes, too, great areas of the surface of the sea are reddened by the vast numbers of small crustacea (Munida), which then play a part of preat importance as food for the fishes and for the guano-producing birds. More striking ‘still are the immense schools of small fishes, the ‘‘anchobetas’’ (Engraulis ringens Jenyns), which are followed by numbers of bonitos and other fishes and by sea lions, while at the same time they are preyed upon by the flocks of cormorants, pelicans, gannets, and other abundant sea birds. It is these birds, however, that offer the most impressive sight. The long files of pelicans, the low-moving black clouds of cormorants, or the rainstorms of plunging gannets probably cannot be equaled in any other part of the world. These birds feed chiefly, almost exclusively, upon the anchobetas. The anchobeta, then, isnotonly . . . the food of the larger fishes, but, as the food of the birds, it is the source from which is derived each year probably a score of thousands of tons of high-grade bird guano. . . . No more forcible testimony to its abundance could be offered than the estimate, made roughly, but with not wide inaccuracy, that a single flock of cormorants observed at the Chincha Islands would consume each year a weight of these fish equal to one-fourth of the entire catch of the fisheries of the United States. As indicated previously, the zonal position of the Humboldt Current fauna is difficult to define. The native organisms have been drawn from various sources outside the area, and the fauna as a whole seems highly nonconforming and eclectic. An analysis of the native sea birds will make this point clearer, but it is also illustrated by other groups of animals. For instance, the shore- fish fauna of the Peruvian coast is mainly temperate rather than tropical in facies. The admixture of sub-tropical with distinctly temperate types of fishes gives the whole assemblage a somewhat Mediterranean cast. The closest faunal affinities of the Peruvian shore-fishes seem to be with the Californias, a considerable number of species being common both to the coast of Peru and to Pacific waters lying just outside the tropics in the northern hemisphere. The distribution agrees closely with the oceanic circulation. Cur- rents flowing from high latitudes toward the equator restrict the Pacific tropi- cal waters to a belt so narrow that it is crossed by a number of species. This belt of tropical water is, nevertheless, a sufficient barrier to divide the shore- fishes of California from those of Peru to a greater extent than those of Florida are separated from those of southern Brazil (Nichols and Murphy, 1922, 513). OCEAN CURRENTS 99 () THE PROVENANCE OF HUMBOLDT CURRENT BIRDS The Humboldt Current is inhabited at one season or another by a variety of ocean birds from many different parts of the North and South Pacific. First, there are the endemic species, known only from waters off the coast between southwestern Ecuador and central Chile. These will receive special consideration below. Second, there are a number of birds of sub-antarctic or pan-antarctic affinities, such as the Magellanic Penguin (Spheniscus magellanicus), the Sooty Shearwater (Puffinus griseus), one or two frigate petrels (Fregetta), the Red-footed Cormor- ant (Phalacrocorax gaimardi), and a tern (Sterna hirundinacea), which extend their breeding ranges northward in varying extent from the Fuegian region to coasts or islands washed by the current. The same is also true of certain shore birds, such as the South American Black Oyster-catcher, and even one land bird (Cinclodes taczanowskii), the ranges of which are scarcely less subject to a mari- time type of contro] than that of the strictly oceanic birds. Third, there is a partially different avifauna in the offshore waters than in the cooler zone close to the land. For example, several petrels of the genus Pterodroma, with nesting grounds on Mas Atierra, Mas Afuera, San Felix, or San Ambrosio Islands, were never encountered by Mr. Beck, during the Brew- ster-Sanford Expedition, until he had gone by schooner a long distance from Peruvian-Chilean shores, and had left far behind the most characteristic birds of the upwelling zone, which is literally within sight of the coast. Fourth, migrants from several different parts of the ocean, some close at hand and some very remote, enter the current seasonally. These include certain species of antarctic and sub-antarctic albatrosses, petrels such as Daption, Prio- cella, Macronectes, Procellaria, and Oceanttes oceanicus, as well as one or more forms of the skua (Catharacta), all of which come from the south. But there are also seasonal migrants from the outer, tropical Pacific, such as the Sooty Tern (Sterna fuscata), examples of which appeared close to shore in central Peru during January, 1925. Furthermore, there are migrants from the Galapagos Islands, and from localities in the tropical waters to northward of the Hum- boldt Current. Among these are the albatross from Hood Island (Diomedea irrorata), a gadfly petrel (Pterodroma phaeopygia), the Blue-faced Booby (Sula dactylatra), the Swallow-tailed Gull (Creagrus furcatus), the Royal Tern (Thalas- Seus maximus), and at rare intervals a tropic-bird (Phaéthon aethereus). Finally, there are large numbers of visitants from the northern hemisphere, including such petrels as Loomelania melania, from breeding grounds off the Lower Cali- fornian coast, gulls like Larus pipixcan, from the interior of North America, and jaegers and phalaropes from the Arctic. We might even add two more categories of birds found in the Humboldt Current, namely, (1) South American continental species, such as a skimmer (Rynchops), a mountain gull (Larus serranus), and a ‘‘zoneless’’ cormorant (Phalacrocorax olivaceus); and (2) birds like the Pale-footed Shearwater (Puffinus carneipes), which come, by a route still unknown, from the Australasian region. 100 OCEANIC BIRDS OF SOUTH AMERICA Only a few species of birds nesting on islands in any part of the Humboldt Current between central Chile and the Gulf of Guayaquil occur also as breeding birds to northward of the region. One of these exceptions is the Blue-footed Booby (Sula nebouxiz), which barely comes into the northern end of the Hum- boldt Current at the Lobos Islands, off the coast of Peru between latitudes 6° and 7° S. Farther northward this booby breeds on tropical islands from Ecuador to Mexico, including the Galapagos group. There can be no doubt about its proper allocation as a tropical sea bird. Diatoms taken from the alimentary tracts of this species (obtained by the birds by way of the herring and other fishes they devour) have proved to belong mainly to warm-water species rather than to the characteristic diatom flora of the Humboldt Current. A second member of this small group of sea birds, which we might call tropical intrusions into the Humboldt Current, is the native pelican. In this case, however, we have at least an endemic subspecies, which ranges along- shore (never going far to sea) from central Chile to northern Peru, where it comes into contact with a tropical race. It should be added here that the American brown pelicans are taxonomically a difficult group, and that the exact relationships of the four or more subspecies occupying the Caribbean- Gulf region, the west coast of North America, the Galapagos Islands, and the Humboldt Current, have never been satisfactorily worked out. In any event Pelecanus occidentalis thagus seems to be a race well differentiated from any of the pan-tropical representatives. The other Humboldt Current species which occur elsewhere as breeding birds, are two small petrels or Mother Carey's chickens (Oceanodroma tethys and Oceanites gracilis). These are also native members of the avifauna of the Gala- pagos Islands which mark, after all, only an outpost of the Humboldt Current, despite the admixture of birds of tropical source. In any event, there has evi- dently been an effective barrier of isolation between the representatives of the two small petrels occupying the Galapagos Islands and the continental islands, respectively, for distinct subspecies of each occur in the two regions. Returning now to the endemic sea birds of the Humboldt Current, ten of which I should call full species and three subspecies, I may offer the following tentative classification relating to the probable geographic origin of the thirteen forms. SOUTHERN DERIVATIVES Or Prospasite NortTHERN ORIGIN Spheniscus humboldti Larus modestus Oceanodroma tethys kelsalli Oceanites gracilis gracilis Pelecanoides garnotii Or Dovustrut SourcE Oceanodroma markhami (north- 5 arya ey Phalacrocorax bougainvillii ern?) : Oceanodroma hornbyi Pan-TrRopicaAL DERIVATIVES Larus belcheri Pelecanus occidentalis thagus Larosterna inca Sula variegata Sterna lorata OCEAN CURRENTS 101 Thus the endemic forms, especially when in combination with the pan- antarctic, pan-tropical, and northern-hemisphere sea birds, which abide in Humboldt Current waters for large parts of the year, give the region a peculiar status. The native birds as a whole do not fit into any of the zonal categories thus far considered, nor is there any analogous avifauna to be found along the eastern coast of South America. Certain resemblances between the west-coast aggregations of littoral birds and littoral fishes are adumbrated. The only choice would seem to be to regard the peculiar and composite Humboldt Current avifauna as of temperate or Mediterranean character, and as zonally distinct from any group of birds previously listed. (2) THE PROVENANCE OF GALAPAGOS SEA BIRDS The Galapagos Archipelago, being closer to the northern hemisphere and to intertropical oceanic islands than the Humboldt Current region, shows an even more miscellaneous composition in its fauna. Incidentally, the same is emi- nently true of Galapagan land birds. In the following list, dealing with the probable origin of the sea birds, the five endemic species are indicated by an asterisk. Or SouTHERN Coot-CurrENT ORIGIN Or CosmopouiTaN Pan-TROPICAL *Spheniscus mendiculus AFFINITIES Oceanodroma tethys tethys Puffinus lherminieri subalaris Oceanites gracilis galapagoensis Sula sula (="‘piscator’’) Sula dactylatra Or InTERTROPICAL PaciFic AFFINITIES Sterna fuscata Pterodroma phaeopygia (from the north) Anoiis stolidus Sula nebouxii (from the east) Or Dovustrut Source Fregata minor (from the west) *Diomedea irrorata (closest to D. ? Or CarIBBEAN AND ATLANTIC AFFINITIES albatrus?) *Nannopterum harrisi *Larus fuliginosus *Creagrus furcatus Oceanodroma castro Phaéthon aethereus Pelecanus occidentalis (subspecies?) Fregata magnificens Eight or more of the above, here indicated by specific names only, have been given subspecific names, and several such races have been regarded as endemic at the Galapagos. g. The Nitto Current. “EI Nifio,’’ the equatorial countercurrent, is a well-known phenomenon off the northern Peruvian coast, commonly appearing about Christmas time or shortly afterwards, and flowing southward. The more or less annual effects of the countercurrent are, as a rule, observ- able only in northernmost Peru. During a longer cycle, traditionally believed to be seven years, its manifestations are more pronounced and extensive. At still longer intervals, representing in the latest instance a rhythm of thirty- 102 OCEANIC BIRDS OF SOUTH AMERICA four years, El Nifio and its associated phenomena may attain their maximum expression. The combined effects during the early months of 1925, for example, were far greater than those of any season since 1891 and probably greater than in any year of record. On January 21, 1925, I measured throughout a ten-hour period a strong, warm current flowing southward, against the prevailing wind, off Point Parifias, Peru. A month later similar conditions still obtained off Point Santa Elena, Ecuador (Murphy, 1926, 31). The immediate result of an advance of El Nifio is to raise the temperature of the littoral ocean water by five or more degrees Centigrade. The normal plankton of the cool Humboldt Current waters next succumbs, perhaps because of the increased temperature, perhaps in part because of a different composition of salts in the water. The common schooling fish leave the region or die, and less familiar species, such as flying fish, dolphins (Coryphaena), and other tropi- cal types, invade the shore waters and even enter harbors. Later, if the incur- sion of tropical waters is marked and widespread, disease attacks the population of cormorants, boobies, pelicans, and other guano birds belonging to the normal Humboldt Current fauna. Carcasses drift ashore in vast numbers, and the sur- vivors of such species are driven southward. Concurrently, tropical sea birds, which do not seem to be affected by the maladies mentioned, follow the movement of El Nifio down the coast into littoral waters ordinarily avoided by them. Man-o’-war birds (Fregata magnifi- cens), a large intertropical booby (Sula dactylatra), and the Red-billed Tropic- bird (Phaéthon aethereus) thus accompany the warm waters to southward of their usual range. Other birds that do likewise include several species of migrants from North America which ordinarily confine their wintering to the equatorial region, such as the Laughing Gull (Larws atricilla) and the Royal Tern (Thalas- seus maximus). Finally, certain antarctic or sub-antarctic birds, such as petrels of the genera Daption, Priocella, Procellaria, and Macronectes, which are normally characteristic of even the northern parts of the Humboldt Current, retreat southward before El Nifio, to be seen no more until the countercurrent cycle has passed (Murphy, 1926, 27). Along the continental coast of the northwestern bight of South America, that is in the permanent Nifio district, there seems to be only one endemic sea bird which does not reach the Galapagos. This is Sula leucogaster etesiaca, a strongly-marked race of the Atlantic Brown Booby. The typical form appar- ently does not cross over from the Caribbean. Farther offshore, as at Malpelo and Cocos Islands, there are one or more other tropical Pacific birds, such as Gygis alba, which do not nest more closely to the cool currents of the equatorial waters or to the continental coast. h. Meteorological, Oceanographic, and Biological Correlations in the Northwestern Bight of South America. The striking climatic and biological changes produced along a tropical shore line by the recurring battles between distinct oceanic zones, as described in OCEAN CURRENTS 103 the last two sections, have been the subject of much recent scientific attention. In a series of papers, Schott (1931, 161, etc.) has interpreted by the dynamic method a wealth of scattered and disjointed observation, including data re- corded during earlier manifestations of El Nifio than the great advance of 1925. The area covered by Professor Schott’s studies is roughly from 30° S. to 10° N. latitude, and from the coast to longitude 100° W., five hundred sea miles west of the Galapagos. The oceanic zones, in a broad sense, are those of the Humboldt Current, the Nifio or Equatorial Countercurrent, and the Pacific North Equatorial Current. It is very interesting to note how surely the included islands serve as indexes of the oceanic environment. Thus the arid Lobos Islands and the Galapagos reflect completely or in part Humboldt Current con- ditions. Cocos, with its luxuriant rain forest, as plainly bespeaks the Equatorial Countercurrent. Furthermore, during “‘Nifio years’’ (of which twelve or more are listed, beginning with 1791), when the terminal front of the southeast trade wind has been pushed back 8°—10° of latitude, heavy rainfall and green vege- tation are as characteristic of the Galapagos as they are of the arid coasts of western Ecuador and northern Peru during the same periods. The Humboldt Current never crosses the equator anywhere near the South American mainland, for the warm countercurrent always flows between it and the Colombia-Panama region. It is therefore impossible that there should be any causal relation between Peruvian conditions and certain recurring meteoro- logical and oceanographic effects in the Gulf of Panama, where the winter lowering of water temperature is due solely to upwelling produced by the period of strongest northeast trade winds. The movement of the Humboldt Current is, of course, likewise dependent upon the prevailing wind trends and the distribution of atmospheric pressures. But the latter are lowest from Janu- ary to April on the Peruvian coast, and lowest during September and October in Panama. These periods correspond with the respective regional rainy seasons. Throughout the whole area under consideration, the oceanic régime is least stable and most complex during the winter months of the northern hemisphere. Thus the Equatorial Countercurrent reaches latitudes south of the equator during February-March but never during August-September. Data from as far as 600 sea miles west of the continental coast demonstrate that the Nifio dis- turbances of 1891 and 1925 were caused by widespread movement of the tropical ocean surface into southern latitudes, which resulted in the forcing of warm and poorly saline waters against the coast to eastward and southward, displacing or overflowing the cooler, saltier waters. Opinion that any part of this vast mass of warm water issued from the Gulfs of Guayaquil or Panama is untenable. During the very period of the Nifio effect the embayed waters just referred to are noted as decreasing in temperature from 4° to as much as 8° C., with an increase in salinity. Indeed, during both 1891 and 1925 a cold current is re- corded as flowing from the Gulf of Panama at the peak of the Nifio disturbance farther southward. It is clear that in the seasons of its notable outbreaks the Nifio movement draws water from the open Pacific—that it amounts to an invasion into the southern hemisphere of the Equatorial Countercurrent. AUG. — SEPT. WATER TEMPERATURE AND CURRENT DISPLACEMENT Miles 250 0 750 Kilometers 0 250 500. 750 1000 30 Fie, 12. [104] FEB.— MAR. \5Callao 19, 4- Legend for maps A,B,C Sis] Warm water 7a Currents 23° Temperature in centigrade 106 OCEANIC BIRDS OF SOUTH AMERICA The distribution of salinity in the eastern tropical area of the Pacific during each northern winter is in itself an indication that the rare Nifio phenomena merely represent an increase in intensity of conditions that prevail annually. Professor Schott marshals his data well for a fundamental explanation. He charts the normal stream axes and borders of the Humboldt and Equatorial Countercurrents for typical periods of August-September, February-March, and for March, 1891, the culmination of the next to the latest great disturbance. The oceanographic circumstances may then be seen to accord with the meteoro- logical background for each period, as outlined above in the meteorological section, under a discussion of the shifting line of doldrums. Whenever the low- pressure belt is drawn south of its normal seasonal position, the southeast trades are replaced by rainy northerly to northwesterly winds, forcing warm surface water southward. A belt of lowest pressure, representing the meteoro- logical equator, and reaching the mainland between 6° and 10° S., was a con- stant factor during the Nifio advance of 1925. Higher pressures prevailed to- ward Panama, thus fulfilling the conditions for northwesterly rain-bearing winds in Peru. It is notable, (1) that the unusual rains always come when such winds occupy the ordinary range of the southeast trades; (2) that they coincide with periods of exceptional wind and rain on the continent; and (3) that the season is always between January and April. The Nifio of 1925 caused the disappearance of typical Humboldt Current phenomena as far south as Arica, Chile. During March alone the rainfall was ten times the total of the preceding ten years! The southward advance of the warm water was slow, but the return to normal was four times as fast, when the trade winds had resumed their course. The dislocation of the meteorological equator, involving the shifting of the equatorial calm belt, with its cloudbursts, thunderstorms, and northerly mon- soons, 300 to 500 kilometers southward, sufficiently indicates that the Nifio results from a widespread atmospheric disturbance. Similar conditions along the west African coast, modified by the absence there of a high mountain system to deflect the trade winds, are referred to below and serve to give ‘‘Nifio”’ a generic meaning in geography. Furthermore, the oscillations of the meteoro- logical equator on the west coast of America have clear analogues in annual conditions during the synchronous northwest monsoon season in the Malay Archipelago and northern Australia. Fics. 12, 13, anp 14. Oceanic conditions off northwestern South America during normal periods of the southern-hemisphere winter season (map A) and summer season (map B), and ata period of exaggerated Nifio phenomena, such as occurred in March, 1891 (map C). The heavy lines numbered 1 and 2 mark confluences of warm waters (chiefly those of the Equatorial Countercurrent) with surface movements to northward and southward. Line 3 is the coolest axis of the Humboldt Current, characterized by active upwelling along the coast; 4 marks the zone of upwelling in the Gulf of Panama during the northeast trade wind season; 5 a zone of upwelling and current divergence contemporaneous with southward displacement of the Equa- torial Countercurrent. The dotted line in map C denotes the approximate course of the meteoro- logical equator at the time of the phenomena indicated. After Schott (1931). MAR. - Year of disturbance 1891 <— Observed current direction <-- Compiled ” ‘ 90 108 OCEANIC BIRDS OF SOUTH AMERICA The significance of the long-term periodicity of these phenomena, which ultimately proceed from varying insolation, hardly comes within the scope of our study. However, Petterson (1929, 121) has pointed out that periods of excessive intrusions of antarctic ice into the southern oceans are always followed by derangement of what we call normal rainfall régimes in the southern con- tinents. It has been estimated that an average of 30,000 cubic kilometers of ice float out of the Antarctic each year and, as Bowman (1930, 445) has said, variation of this on a colossal scale could not fail to have profound and far- reaching after-effects. In the year 1922, according to Petterson, floes and bergs extended outward so far from the Antarctic that for safety’s sake the southern trade routes had to be shifted northward. The outbursts of ice were succeeded after an appro- priate lag by notable movements of tropical surface waters into the southern hemisphere. The Humboldt-Nifio conditions of 1925 were paralleled, for example, by a simultaneous deflection of the cool west African Benguela Current, permitting warm waters from the Gulf of Guinea to expand southward, and bringing tor- rential rains, corresponding with the inundations of arid western South America. Nyasaland and Rhodesia received the greatest rainfall experienced within European knowledge. At Mossamedes, latitude 15° 11’ S., perhaps the most arid station on the coast of west Africa, the rainfall in February, 1925, was more than double the highest record previously known for that month. From a marine zodlogist’s point of view, the advantages of such a time of geographic crisis are that they enable him to see, through the medium of a drama compressed within a single season, an epitome of the secular process which has been going on steadily if less violently throughout the ages. In the northern waters of the Humboldt Current, for example, he finds many animals beautifully adapted to a special and peculiar environment, and he can compre- hend to some extent the selective action of this environment in molding the organism to a “‘fitness’’ which acquires ironic significance as soon as the en- vironment itself alters. He cannot but be impressed also by the delicate balance of the controls over the environment. One might almost say that the huge and highly specialized aggregation of marine plants and animals in this region have but one danger to fear, namely warm water. Because of such reasons the oceanic bight bounded by northern Peru, the coast to northward, the Gala- pagos, Cocos and Malpelo Islands, offers one of the most fruitful fields for a study of the forces and inhibitions that determine the distribution of life in the sea. (1) ZONELESS BIRDS In a previous discussion of birds common to the Antarctic and Sub-Antarctic Zones of the South American region, reference has been made to the extraor- dinarily wide climatic scope in the breeding ranges of certain coastal and in- sular species (p. 76). Thus a tern (Sterna vittata), which appears to be a close but mainly sedentary relative of the Arctic Tern (Sterna paradisaea), nests in the ZONELESS BIRDS 109 Antarctic Archipelago, at the South Orkneys, South Georgia, Gough Island, and Tristan da Cunha. The common allegation that it also inhabits St. Helena and Ascension proves to be erroneous. Nevertheless, the one species, no doubt represented by two or more subspecies, resides both at icebound islets of West Antarctica and at Tristan, the latter being at the extreme northerly edge of the Sub-Antarctic Zone. In this case water temperatures and zonal convergences seem to be thoroughly disregarded, although further taxonomic study may throw new light upon the subject. Another tern (Sterna hirundinacea), which is much more satisfactorily known and which has not been subdivided, breeds at the Falklands and at Cape Horn, from these latitudes northward along the east coast of South America to the neighborhood of Cape Frio, Brazil, and on the west coast at least to central Peru. The widely accepted belief that it nests also at some of the antarctic islands proves, however, to be without foundation, and to be due to confusion of the species with Sterna vittata. Sterna hirundinacea is, in fact, confined to the continental littoral. It is typically sub-antarctic, and might appropriately have been listed above among the birds of that zone of surface water but for the fact that its range extends far into the so-called tropics because of the influence of cool currents and cool belts of upwelling. It is, therefore, not properly a ‘‘zoneless’’ bird, but is discussed here because of the length of its distribution across the parallels of latitude, and because of its undeserved repu- tation of being both an ‘‘antarctic’’ and a “‘tropical’’ sea bird. The Kelp Gull (Larus dominicanus) is more notable. It nests at the South Shetlands, South Orkneys, South Georgia, in the Falkland-Magellanic district, northward to Rio de Janeiro, and to the Lobos Islands off northern Peru. One of its centers of maximum abundance as a breeding bird, indeed, is at Lobos de Tierra (6° 30’ S.), well over fifty degrees of latitude north of its polar nesting grounds. Terns subsist upon both small fish and swimming crustaceans, which in one form or another must be available at all the breeding stations throughout the extensive ranges mentioned. The Kelp Gull has a broader taste in food, making use of carrion and offal, fish, shellfish and other beach invertebrates, as well as the eggs and young of many kinds of birds. In fact, its principal foci of popu- lation, whether in frozen or in torrid localities, are usually closely associated with colonies of weaker species upon which it preys. Even taking all these matters into consideration, such amplitude of range as that presented by the Kelp Gull is a distinctly oceanic phenomenon of distribution, without any equivalent in the northern or “‘continental’’ hemisphere. Still more remarkable than any of the foregoing examples, is the case of an American cormorant (Phalacrocorax olivaceus). The specific range is both inland and coastal from Cape Horn to the southern United States, but the birds occur- ring north of Nicaragua have been separated as subspecies mexicanus, while a third race seems to inhabit Tierra del Fuego. The typical South American form disregards not only latitude and climate, but also altitude, water temperature and salinity, precipitation, the nature or presence of vegetational ground-cover, 110 OCEANIC BIRDS OF SOUTH AMERICA and almost every other obvious environmental factor. It lives and breeds along rocky, sandy, or muddy seacoasts. It is equally at home beside a glacier in southern Chile, on a barren rainless islet off Peru, among mangrove channels of the Caribbean coast, in the stump-filled lakes of Panama, on tepid rivers in the forested interior of the continent, or in lofty and frigid mountain lakes such as Titicaca and Junin. It is clear that this cormorant is superior to climate, in the ordinary sense, and that it could not possibly be fitted into any of the zones we have discussed. The probable explanation of its freedom from environ- mental control, aside from its patent toleration of a wide range of heat and cold, is that it feeds upon almost any sort of bottom-living fish, of which there is no dearth in either salt or fresh water of various latitudes, temperatures, and composition. AN ORNITHOLOGICAL CIRCUMNAVIGATION OF SOUTH AMERICA L' acqua ch’io prendo gia mai non si corse—DANTE The journey upon which we are embarking has never before been undertaken. I propose to skirt the continental coast line, with full disregard of distance, time, and circumstance. Neither weather nor current need retard us; no island of the surrounding seas shall be too remote for us to reach within a twinkling; moreover, we may view the successive prospects, as preference dictates, either through eyes of today or those of generations long departed. To describe the topography in detail would lead us too far afield. Let us in the main travel fast, halting only to picture a series of well-chosen localities along the ocean front, and at islands offshore. At each such station we may select the time of year best suited to our purpose of observing the bird life or the periodic natural phenomena concerned with the distribution of birds. 1. Tue CarisppeEaAnN Coast We may begin our survey of the South American coast at the Caribbean end of the boundary between Panama and Colombia (8° 41’ N.), where the promon- tory of Tiburén, the ‘Cape of Sharks,’ marks the northwestern extremity of the Gulf of Uraba. Southeastward as far as the low and relatively sheltered projection of the Atrato delta, the shore is high and bold, and a wild sea beats upon it during the windy season of the northern-hemisphere winter months, which is also the dry season. Inland and to southward, these same winds pre- cipitate rain, but, as noted heretofore, coasts directly to leeward of strong trades are prevailingly dry. Close to shore, along the western side of the gulf, are numerous steep islets, and, beyond the cliffs, the rtdge of the Espiritu Santo range, rising to two thousand meters or more, can be seen clearly from the water. About the mouths of the River Atrato a foreshore has built out, and the high coast here gives way to sandy, mangrove-covered cays which line the whole head of the Gulf of Uraba, the almost fiord-like body of water extending southward from the broad bight of the Gulf of Darien. Along the western shore of the inner gulf CARIBBEAN COAST HT - 60 “Bermuda a | ff ; / s New Orleans Le Jacksonville rf 30 pong —- a 2 rae Co Ea i / \) a z / - . | = 8) . a «av / me f 1 apo} MS INe / BN, / c = % ¢ / Sa OR t aa ! I. de Pinos t ils ee oy 64 Cayman Belize] 0 o& S \xAntigue Aeuadeloupe S. Juan del Norte 3 a eos: Oe ad@ x 35 FeinidadG "J PtoLimén Cyman: DOR 10 R. Orinoco ony Geor getowryro. Pan R Essequebo ~~ R.Courontyne’ Rharon ™ 60 Cayenne Beck's route aS Murphy's route 30 Fic. 15. Coasts of northern South America, the Caribbean region, and the Gulf of Mexico. In addition to the indicated courses of expeditions by Beck and Murphy, figures in circles show the location of other American Museum field work, as follows: 1, Thayer, at Grenada; 2, Tate, at Cumana; 3, Chapman, at Trinidad. For details of the Venezuelan coastal islands, see Fig. 16. the Atrato pours forth through thirteen mouths. Flowing as it does from a basin of almost perpetual rainfall, it carries a tremendous quantity of water, despite its insignificant appearance on small-scale maps. Indeed, the pilot books of the South American coast credit it, though perhaps incorrectly, with being fourth in volume among the streams of a continent which is the mother of great rivers. Eight mouths of the Atrato are navigable for small craft, such as native cargo sloops and the bongos or dugouts that are common to all forested coasts of northern South America. Delta deposits extend far into the gulf, and bars blocking the mouths show depths of two meters or less, a condition hardly alleviated by the tidal rise, which is very slight. Inside the delta, however, the river assumes impressive proportions, with soundings in the channels of from 10 to 20 meters, and a breadth sometimes exceeding a kilometer. Below Sucio, 100 kilometers from the mouths, there are no permanent habitations on the low banks, for the surrounding country is inundated during ten months of the year. The current of the river is swift, the overflow through- out widespread adjacent territory serving to relieve the head of flood-water that accumulates during the rainiest season. The eastern shore of the Gulf of Uraba is flat and swampy, with muddy 112 OCEANIC BIRDS OF SOUTH AMERICA shallows for a long distance off. Toward the northerly end of this coast is a low sandy peninsula, with the large Aguila Lagoon behind it, and still north- ward, protected by extensive shoals, Point Caribana marks the eastern gatepost. As already observed, the months of calm in trade-wind regions are also the months of rain. In the northern springtime, when the northeast trade wind withdraws its distal fringe so that the Caribbean coasts of Colombia and Venezuela no longer submit to its influence, the rainy season commences, to continue until September or October. The amount of rainfall differs markedly within short distances along continuous coast lines and, in most localities of even heavy precipitation, there is usually a rhythmic concentration of rain during certain hours, leaving other parts of the day or night relatively clear. The following description of weather conditions in the Gulf of Urabé is from the report of Michler, whose vessel remained in the gulf for four months including, however, only the beginning of the rainy season, during the earliest United States surveys for the project of a ship canal between the Atlantic and Pacific. The days and nights during this time were beautifully bright; in but few climes can the gaze rest upon the distant splendor of more perfectly glittering star-light skies. . . . The weather was usually clear. . . . It rained only once by day, though very slightly, and but once for a few moments after dark. So far from being the horrible climate represented by others, one of continuous storms, excessive heat, and miasmatic atmosphere, the veritable Pandora’s box of all the malaria of the most feverish portions of the world, the atmosphere proved to be extremely delicious. It so continued during the entire stay of the Varina in the Gulf, from the first of December to the latter part of March, with the exception of two nights previous to weighing anchor, homeward bound. As if not to let her depart without a benefit, heavy rains, such as are only known in the tropics, fell for hours upon her decks, attended by most terrific peals of thunder. The play of lightning was so quick and intense as to vivify all surrounding nature. According to the most reliable information the rainy season along the Gulf extends from early in April to the latter part of November. The rains are not, however, by any means excessive even during this period, but occur chiefly in the shape of short, smart showers of from a few minutes to some hours duration, especially during the night, with occasional heavy and prolonged falls of one or two days. It is even by no means uncommon for intervals of from three to six days to elapse during the wet season without a day of rain. Subsequent experience of some months along the valley of the Atrato and San Juan proved that by far the greater portion of the rain of that region falls during the night (Michler, 1861, 28). Such meteorological conditions are altogether typical of many coastal sec- tions of northern tropical South America, whether they face the Caribbean, the Pacific, or the Atlantic. With the exception that the annual period of the rainy season would be different, the description from Michler might be equally well applied to weather in the Gulf of Guayaquil or in the estuary of the Amazon. As with climate, so largely with the plant life of the shore, as well as with the bulk of the beach and marsh birds. From northernmost Peru around the northern edge of the continent to the Amazon or beyond, the plan for both coastal vegetation and coastal bird life is a common one, more or less modified by local influences. For the most part, the same herons, ibises, storks, spoon- bills, and anhingas roost among the same species of mangrove and courida, while overhead soar the same forms of the Black and Turkey Vultures. The CARIBBEAN COAST 113 same resident or migrant Limicolae patter along the wash of the waves or wade in the shallow lagoons; the same terns and skimmers quarter the brown estuaries; the same cormorants, pelicans, boobies, and man-o'-war birds pursue schooling fish to varying distances offshore. There are, of course, changes and exceptions. Here a familiar species will be squeezed out by some inhibitory influence of the environment; there a new intrusion may appear; again, the range of a certain bird may end as the form becomes replaced by a representative subspecies or species. But, in the main, the avifauna of the tropical coasts is a common one, witha continental, rather than an oceanic, element predominating. The boobies and man-o’-war birds and, to a lesser extent the terns and pelicans, reach the shores from outlying islets: to find the source of the tropic-birds and the few petrels that visit such coasts, we should have to go still farther afield. With reference to coastal birds, and factors determining their distribution in the Santa Marta district, Todd and Carriker (1922, 67) write the following, which has a bearing upon what I have said of the tropical water birds in general: The Littoral is characterized by the great diversity of habitats within its limits, each of which has its peculiar species of birds. (There are some forms, however, which are present in nearly all situations.) . . . As already explained, this diversity of conditions is due in the main to the topography of the region, differences in the relative humidity, etc. Beginning in the west, we have the mangrove-lined shores and waterways of the Ciénaga Grande and Magdalena delta, inhabited by many species of aquatic birds, some of which are rare or absent elsewhere in this general region. . . . Most of the water birds . . . havea more or less extended distribution in tropical America, and their presence depends mainly upon suitable habitat and local conditions, not upon zone. Rainfall, as an example of such local factors, varies through very wide extremes along the section of the South American coast under consideration in this chapter. The head of the Gulf of Uraba lies within the same climatic dis- trict that rules the Pacific coast of Colombia, with an annual precipitation of over 2000 millimeters. But along the Caribbean coast, stretching in a north- easterly direction from the gulf, less and less rain is received toward the tip of the Goajira Peninsula, except in a narrow seaward-facing strip between the Sierra de Santa Marta and the sea. The remainder of this stretch lies in a moun- tain “‘rain-shadow,”’ a familiar phenomenon on many restricted parts of the South American shore line. Beyond the Santa Marta range, in Goajiros, rainfall rapidly gives out. The isohyet of 500 millimeters encloses a broad belt of arid coastal country on both sides of the Gulf of Venezuela, including the whole of the peninsulas of Goajira and Paraguana. To eastward of Puerto Cabello, on the Gulf of Triste, the line runs alongshore, close to the Caribbean, as far as the Gulf of Cariaco. Both Margarita Island and the lower parts of the Paria Peninsula thus lie in the zone of prevailing aridity which, as Dr. Frank M. Chapman tells me, ends on Monos Islet, in the Bocas de Dragos, without quite reaching the island of Trinidad. Moreover, the composition of the Caribbean coastal water varies consider- ably, owing to the direction of surface movements in combination with the increment from great rivers and the strength of the trade wind. From the Gulf 114 OCEANIC: BIRDS OF SOUTH AMERICA of Uraba to the delta of the River Magdalena the coastwise current sets chiefly in a northerly or northeasterly direction. The region is ‘‘around the corner’’ from the general westward-setting movement of the open Caribbean, and the local flow is doubtless part of the countercurrent system that runs in a southerly direction along the Central American coast and loops to eastward in the Gulf of Darien. Off the mouth of the Magdalena, and the coast of eastern Colombia, the main Caribbean or equatorial current is in evidence. A rip or race often forms where the outpouring river water and the ocean current come into con- tact, an effect which has been observed as far as a hundred kilometers to north- ward of the delta. Here the water is bluer than in the eastern Caribbean, and sometimes long rows of floating seaweed, teeming with animal life, mark the lines between deep ocean water and the freshets from the land (Paul, 1932, 142). The Atlantic Equatorial Current enters the Caribbean between Trinidad and the southerly islands of the Lesser Antillean chain, its strength being greatest in the strait between Trinidad and Grenada. Part of its influence is also appar- ent, as will be noted Jater, in the narrow openings into the Gulf of Paria, be- tween Trinidad and the Venezuelan mainland. Inside the Caribbean, the current sweeps mainly north of the Venezuelan coastal islands, but, during periods of strong trades, surface tongues flow southward between these islands and form eddies along the continental coast. In the extreme southwest part of the Caribbean Sea the Equatorial Current can never be apparent, although the trade wind sometimes blows its clear and saline surface water into the Gulf of Darien. This wind, and the ‘‘vendavales’’ or westerly land winds of the region, together with the local movements of the surface for which they are responsible, occasionally bring about a concentration of driftwood, from many rivers and from various shores of the Spanish Main, in the constricted Gulf of Uraba. The mighty Magdalena is the second 6f the great trio of rivers on the north- ern slope of the continent, the Orinoco being the third. The Magdalena enters the Caribbean through one gigantic main mouth, but many other exits also cut through the mangrove-covered flats along an extensive stretch of shore line. The distance between the most easterly mouth, the Boca de la Ciénaga, which flows northward into the Caribbean, and the westward-opening Boca de Ma- tuna, south of Cartagena, is about 160 kilometers as the crow flies. Along part of the distance between these, a line of low coastal hills has forced the drainage to take several courses toward the sea. The upper delta of the Magdalena, the area now occupied by the Ciénaga Grande, together with the network of lagoons between it and the river, was within recent geological time an arm of the Caribbean, like the present basin of Lake Maracaibo. It is evident also that this area, as well as the littoral to eastward, has been uplifted at no very remote period. The steady supply of sediment carried down by the Magdalena and the rivers from the westward slopes of the Sierra Nevada are gradually filling up the brackish Ciénaga Grande. Mangroves have secured a foothold on all sides so that there are now hundreds of square kilometers of mangrove swamps, interspersed with endless waterways that afford food and cover to myriads of aquatic birds (Todd and Carriker, 1922, 16). CARIBBEAN ‘COAST 115 The shallow bars that make the mouths of the Magdalena impassable to large vessels are covered by deeper water during the dry season than the wet, because of the vast load of silt deposited throughout the latter period. In freshets the current of the main stream may run as fast as 11 kilometers per hour. A rolling and treacherous sea is rarely lacking on the outer bar. Much driftwood, including large treet runks, is transported by this route into the Caribbean. From the Gulf of Uraba to the “‘rain-shadow”’ strip east of the mouth of the Magdalena, most of the islets off the shore are wooded, some of them show- ing groves of royal palms surrounded by other vegetation. But eastward from the Goajira Peninsula and the Coro region of Venezuela the entire coast line, as well as the outlying chain of islands, has a generally parched appearance as far as the heights of the Paria Peninsula and Trinidad. Just east of the Mag- dalena delta, the magnificent snow-capped Santa Marta Mountains, with their highest peaks nearly always above the clouds, are visible from sea. They are, of course, responsible for the greater precipitation and increased luxuriance of the foreshore between Cape Aguja and Point Caricari. Wollaston, en route for the Sierra, journied westward from Rio Hacha to Dibulla. He writes: Riding along the beach is not so simple as it sounds. The rise and fall of the tide is a few feet only, and the beach is a narrow strip of soft sand with the heavy surf of the Caribbean Sea on one hand and mangrove swamps or muddy lagoons on the other. In many places stranded logs or long projecting mangrove branches make a fence across the sand and compel the reluctant mules to take to the Caribbean Sea (Wollaston, 1925, 97). Turning southward west of Dibulla toward the mountains, Wollaston’s party encountered first a tract of grass and thorny bush, interspersed with swamps, where gigantic horseflies and other insects made travel almost intoler- able until the coolness of the wooded foothills was gained. The territory on either side of the Gulf of Venezuela, or Maracaibo (the northern tip of South America), is the driest coastal region on the whole northerly face of the continent. Inland, on the shores of Lake Maracaibo, and nearer the sea on the coast to eastward of the Gulf of Triste, the precipitation is by no means scanty but, as noted above, the peninsulas of Goajira and Para- guana receive less than 500 millimeters of rainfall annually, while the Coro region, at the base of the Paraguana Peninsula, has not more than 250 milli- meters. The outer gulf, and the inner, land-locked Lake of Maracaibo, therefore coincide with a north-south section extending from districts of extreme aridity to those of ample rainfall. The water of the so-called lake varies from a brackish solution, near its narrow and shallow connection with the gulf, to water com- pletely fresh toward its inner end. The region should offer a field of particular interest for a study of water birds, but apparently little ornithological work has been undertaken or, at any rate, reported upon. Platt (1934, 157) describes the Maracaibo Basin as seen with the mind's eye from an airplane in which he had flown eastward from Colombia across the Sierra de Perija. From the air the whole bowl-like formation seems perceptible —the nearly encircling rim of mountains, the belt of lowlands, and the central 116 OCEANIC BIRDS OF SOUTH AMERICA body of water. Dense forest on the slopes, broken only by streams, gives way at low levels to scrub woods, spotted with clearings which are increasingly numerous toward the lake. These openings indicate “‘haticos”’ or little ranches, each with an area partly cleared, a patch of cultivation, a small house, and a water-hole. The district has aspects of semi-aridity consistent with an annual rainfall of about 508 millimeters, falling during a season of seven months, average temperature of over 27° C. for every month, and rapid drainage through the sandy soils of partially consolidated coastal plain sediments. This combi- nation does not necessarily mean desolation. Mistletoe is as plentiful as cactus, and live animals are more in evidence than dead ones. The principal market- able products of such districts are goatskins and dividivi, the latter being pods of the prevalent leguminous tree of the area, valuable for their content of tannic acid. At Santa Rosa, an ancient lake village on the neck of water joining Lake Maracaibo with the Gulf of Venezuela, continues Platt, goats and dividivi are almost if not quite supplanted by fish and coconuts. Here the lake water is brackish, and fresh water is to be obtained only at some distance inland from holes dug in sand. But the distinctive form of settlement, with dwellings on piles, such’as originally gave the name of Venezuela, or Little Venice, to the country, is one befitting local circumstances here where the shore is low, partly swampy, and covered with a tropical tangle ordinarily called jungle. The water of the lake and strait are calm and shallow, and simple pile construction provides a more healthful home on water than could be established on land, with relative security from insect and other enemies. On the eastern shore of Lake Maracaibo, as about Lagunillos, a sinking of the land and the margins of the lake bottom, attributed to the recent extraction of petroleum, is causing rapid flooding of large areas. The topography of the floor of the sea north of Venezuela conforms more or less with that of the mainland. Parallel with the coastal cordillera is the sub- merged ridge upon which the Leeward Islands lie. That these fully share the coastal aridity is evidenced by the following figures on total annual rainfall, from Franze: Average Maximum Minimum Aruba 438 942 85 Curagao 559 1105 265 Bonaire 440 870 164 Margarita 245 The even greater dryness of the continental shore line is shown by the annual mean precipitation for Coro, given above, that for La Guaira (283 millimeters) and for Cumana (258 millimeters). The bulk of the rainfall on both mainland and islands comes between the months of October and January, the dry season of northeast trades commencing during the latter month and continuing until July. While many of the above means are based upon figures covering only a few years, they agree well with comparable records for periods of thirty years or more. GARTB BEAN aGoasT 117 During the wet season on the Caribbean coast, the country adjacent to the sea is often foggy in early morning, but by sunrise the land breeze (terral) begins, blowing sometimes very strongly through the forenoon. The afternoons are apt to be calm and fair, although a haze, said by the pilot books to resemble the west African harmattan (but nowise akin to it), overhangs the coast much of the time. The seasonal retraction and extension of the trade-wind border are due, of course, to the swing of the belt of greatest insolation toward and then away from the tropic of Cancer. Toward the end of the northern summer, the region of transient calms and rainfall is gradually reinvaded. From October to April the trade wind blows throughout the day on the shores of Venezuela, to be in- terrupted nightly by the terral, which is as characteristic of evening during the windy season as it is of morning during the calm season. At the larger islands off the coast there is likewise an alternating régime of land and sea breezes, the nocturnal wind blowing most strongly on the coast which is in the lee of the trades, in other words, where the terral is reinforced by the trade wind. To westward of the Goajira Peninsula, the trade winds are exceedingly strong, sometimes amounting to gales. The Spanish sailors of early days called these impetuous trades, which differed so markedly from the gentler and more constant winds of the open oceans, ‘‘los brisotes de Santa Marta’’ (von Hum- boldt, 1852, 1, 127). During the windiest period (January to April), the source of the breeze on this westerly-facing coast lies generally in the west in early morning, later veering through northwest and north, and by noon blowing parallel with the shore line. Such ‘‘following of the sun”’ is a tendency of all regular breezes like the trade winds, which should by no means be thought of as keeping a steady northeasterly course at all localities and during all seasons. The trade winds are readily deflected by the meteorological system of a con- tinental coast, and as a rule they tend to blow perpendicularly toward the shore when the latter is in the direction of the sun. The same principle applies to the annual as well as to the daily cycle, and many correlations can be observed in the West Indies. For example, winds from north of east prevail from November to March; April and May, the most tranquil months of the year, are characterized by easterly breezes; winds from south of east prevail from June to October. The latter part of the summer is a time of feeble trade winds, leading to the hurricane months of August, Septem. ber and October. As explained heretofore, no part of the continent of South America lies within the path of hurricanes, although these storms frequently sweep through the length of the Caribbean, parallel with the Venezuelan coast, spreading devastation from the Lesser Antilles to Yucatan. The South American main- land, however, as well as Trinidad and the chain of coastal islands, is all but free from their direct effects. Much further information is needed about the “ornithology”’ of Caribbean hurricanes, particularly with regard to the boobies, frigate-birds, flamingoes, song birds, etc., sometimes vaguely reported by newspapers as being seen buffeted about near a ship in the vortex. 118 OCEANIC BIRDS OF SOUTH AMERICA In the central part of the Venezuelan coast, near Puerto Cabello on the Gulf of Triste, the mountains turn northward toward the sea. Thence eastward the range hugs the shore, and at many places toward Point Pefias on the Paria Peninsula or, indeed, quite to the northeastern tip of Trinidad, the waves of the Caribbean break against the foot of the mountains themselves. The green aspect of the heights is due to constant precipitation which does not affect the aridity of the lowlands except where streams descend through forested ravines to sea level. Between La Guaira and the Gulf of Cariaco much of the immediate shore is low and dry (save for mangrove swamps). Of this district, which 1s reminiscent of the coast of Peru, Robinson and Lyon (1902, 136) write: This northern coast of Venezuela is noted for its heat. The tropical sun beats upon it without mercy, and where water fails the aspect of the country is that of a desert. The littoral plain near La Guaira, where not watered by the acequias, or irrigation trenches, which bring the water from high up on the courses of the mountain streams, is parched and dusty, and given over to Agaves, Cacti of various kinds (Opuntia, Cereus, Melocactus), and to a sagebrush-like plant (Lantana) of characteristic odor. But wherever water can be obtained the plain assumes a most fertile and flourishing character. Such is the condition at Macuto, three miles east of La Guaira. Still farther to the east, some seven miles from La Guaira, a large stream comes down a fair-sized valley. Before debouching upon the plain, its water is all carried off to the right and left by the acequias, and used to operate sugar mills and afterwards to irrigate wide fields of cane. Close to shore in this broad bight of the coast are eleven small islands, com- prising two clusters, which were passed and described by von Humboldt and Bonpland during a voyage from Cumana to Caracas at the end of the eighteenth century. They write: We found ourselves at midnight between some barren and rocky islands, which uprise like bastions in the middle of the sea, and form the group of the Caracas and Chimanas. The moon was above the horizon, and lighted up these cleft rocks which are bare of vegetation and of fan- tastic aspect. The sea here forms a sort of bay, a slight inward curve of the land between Cumana and Cape Codera. The islets of Picua, Picuita, Caracas, and Boracha, appear like fragments of the ancient coast, which stretches from Bordones in the same direction east and west. . . . Several of the islands are visible at Cumand, from the terraces of the houses, and they produce, according to the superposition of layers of air more or less heated, the most singular effects of suspension and mirage. . . . As we came near this group of mountainous islands, we were becalmed; and at sunrise, small currents drifted us toward Boracha, the largest of them. As the rocks rise nearly perpendicular, the shore is abrupt; and in a subsequent voyage I saw frigates at anchor almost touching the land. The temperature of the atmosphere became sensibly higher whilst we were sailing among the islands of this little archipelago. The rocks, heated during the day, throw out at night, by radia- tion, a part of the heat absorbed. As the sun arose on the horizon, the rugged mountains projected their vast shadows on the surface of the ocean. The flamingoes began to fish in places where they found in a creek calcareous rocks bordered by a narrow beach. All of these islands are now entirely uninhabited; but upon one of the Caracas are found wild goats of large size, brown, and extremely swift (von Humboldt, 1852, 365). Of the Gulf of Cariaco, and the country on either side of it, Lowe (1909, 305, 321) writes: To one steaming up the Gulf of Cariaco nothing indeed could be much more striking than the strange contrast which is exhibited on the one hand by the Peninsula, with its arid, desolate, and schistose rocks only scantily clothed with cactus and mimosa scrub, and on the other by range CARIBBEAN COAST 119 after range of tall mountains, Juxuriantly forest-clad, which mark the Secondary limestone forma- tions of the mainland. . . . Once inside, the voyager finds himself on an irregular sheet of deep still water of the deepest blue, studded here and there with small islets. Surrounding it on all sides, in the fashion of a Scotch loch, are high hills. They present a series of colours which vary from rich deep red to deep yellow-ochre. In the distance the iron-tinted soil seems in places to be frosted with old silver, an effect produced by the patches of silvery-toned cactus-scrub. Along the shores, stretches of bright golden sand alternate with.the deep green of the mangrove-belts. Here and there on the lower slopes patches of acacia-trees and the candelabra-like inflorescence of the agaves relieve the more sober tones of the cactus-scrub. When von Humboldt embarked at the town of Cariaco, in late September, 1799, for a boat journey to Cumana, he encountered a morning torrent of rain, accompanied by lightning and heavy thunder. “‘Swarms of flamingoes, egrets, and cormorants,’’ he wrote (1852, 289), ‘‘filled the air, seeking the shore, whilst the alcatras, a large species of pelican, alone continued peaceably to fish in the middle of the gulf.”’ The mountain ridge forming both the Paria Peninsula and the backbone of Trinidad, against which the full strength of trade wind surf beats, is breached by the Bocas de Dragos. Issuing here into the Caribbean is a powerful current made up of Orinoco waters combined with that part of the equatorial ocean stream that has found a way through the funnel of the Serpent’s Mouth into the Gulf of Paria. The tidal range in the gulf is greater than in adjacent parts of the coast, and during the rainy season the outflowing current, augmented by an ebbing tide, sometimes attains a velocity of about 10 kilometers per hour. The difficulties of the islet-studded passage were commemorated by Columbus in the name which it still bears (‘Dragon's Mouths’), but with the advent of steamers the term has lost some of its old significance. Here, at the extreme eastern end of the Caribbean coast, there begins to be a more copious rainfall at sea level than is found in northern Colombia and Venezuela. .The northeast trade wind season holds from November to May, the wet season during the remainder of the year, with the heaviest precipi- tation in July and August. At Trinidad, May is the warmest month of the year, with a mean temperature of 33.5° C., and February the coolest, with 19.4° C. The occasional remarks made on the Caribbean mainland water birds and shore birds, in this chapter, may be concluded with a brief record of Mr. G. H. H. Tate's observations at Cumana, Venezuela. Between May 16 and 24, 1925, after an American Museum expedition to Mt. Turumiquire, Tate made headquarters at the port of Cumana, which is separated from the main town by a stretch of sandy, salt-encrusted plain. One morning he set forth long before dawn in a rowboat, with the object of reaching some extensive growths of swamp vegetation to eastward of the port by the hour that the marsh birds would leave their roosting places for the feeding grounds. Approaching mangrove-covered Punta Erga just after day- break, he could see the pale pinkish glow of a flock of Roseate Spoonbills in the earliest rays of the sun, together with a scattering of egrets and several 120 OCEANIC BIRDS OF SOUTH AMERICA other sorts of herons. Vast numbers of plovers and sandpipers were feeding on expanses of mud flats left bare by the tide. Flocks of skimmers were already coursing up and down the channels. Toward solid land, a wall of taller man- grove trees evidently served as a nightly roosting place for more herons and for hundreds of cormorants. As Tate stepped out of his boat, to wade shore- ward through alternating stretches of mud and shallow water, groups of gor- geous Scarlet Ibises flew out of the’ low mangrove thickets and circled around the visitors before disappearing behind the trees. Mr. Tate collected in the territory close to Cumana about twenty-five species of water and shore birds. The Limicolae naturally included many migrants, presumably bound toward North America, such as phalaropes, turnstones, curlews, stilts, willets, knots, plovers and sandpipers, in addition to such resident forms as the tropical American Oyster-catcher. Local birds belonging to a more distinctly salt-water aggregation included the Brown Pelican, Man- o'-war Bird, Least Tern, and the estuarine Large-billed Tern (Phaetusa). He noted that birds of the last-named species often circled at night around the powerful electric light on the pier of the port. The Bigiia Cormorant occurred, as might be expected, in every sort of damp habitat—fresh, brackish, and salt, forested or exposed. Such birds as those listed are substantially what one would expect to see in suitable places along the entire length of the Caribbean mainland shore, from Trinidad to Panama. The Man-o’-war Bird is the only one that might not normally penetrate far into fresh-water districts, such as the Lake of Maracaibo. However, the presence of a fortuitous source of food, such as a slaughterhouse on the shore, or a place where fish are gutted in large quantity, is sufficient to draw Man-o'-war Birds from their usual salt-water foraging grounds a long distance up inland waterways, including even forested rivers. Strictly marine birds, like boobies, can be observed from the mainland shore only in the straits between the continent and their nesting islands. To west- ward of the Goajira Peninsula, and particularly along the jungle-bordered coasts of the Gulf of Darien, oceanic birds of all sorts are scarce. The ubiquitous cormorant is present, however, not only in the flats and swamps of Uraba, but even along the Stygian course of the Atrato. According to Todd and Carriker (1922, 131), the Brown Pelicans probably nest on some of the remote insular beaches between the mouth of the Magdalena and the town of San Juan de Ciénaga, Colombia. Turning now to the chain of Caribbean islands, which are the main center for the oceanic fowl of this littoral region, it must be pointed out that the conditions for bird life have naturally altered enormously during several cen- turies of European occupancy or interference. Even on the islets that have never been the permanent home of men, the introduction of new plants and of domes- tic animals has changed the pristine state to an extent hard to conceive. The present appearance of some of these islands would no doubt put them beyond the recognition of one who had known them only in primitive times. The changes referred to include many phases of human activity, from the removal CARIBBEAN COAST. 121 VENEZUELAN COASTAL ISLANDS Miles 100 iso 200 250 Kilometers ° 50 100 so 200 «4250 ° w ° Heights in meters iside Aves 2%R h Los Roques | 83 Oo. -2e83 . Caracas 1942 E W ue Z U i IL A Fic. 16. of old guano beds from such sites as Little Curagao and El] Roque to the estab- lishment of ostrich farms on Aruba! The larger and higher islands epitomize the opposite mainland coast, with xerophytic plant life and semi-desert types of birds near sea level, and tropical rain forest, with a correlated avifauna, near the summits. The rainfall in high altitudes is to be associated with adiabatic cooling of the sea winds, exactly as on the continent. Mountain torrents are numerous, combining to form streams that reach the sea, at least during the rainy season. Margarita, the largest member of the chain, is thus described, with its land birds, by Clark (1902, 259): The island has three well defined life zones; first, the flat and hot coast region, sandy, and with scant vegetation, consisting of post and melon cacti, with the dreaded ‘tufia’ and thorn-trees. This extends all around the island, and occupies a strip averaging about three miles in width, in which the Burrowing Owl (Speotyto) is exclusively found, while the Troupial (Icterus), Scaled Dove (Scardafella), Buzzard (Buteo), and Parrakeet (Conurus) prefer it to any other region. The chief towns of the island, Asuncién, Juangriego, and Porlamar, are situated here. Next comes the intermediate region, of rough, hilly country, with a large amount of scrubby growth, and many varieties of cactus, forming the home of the Tawny Cuckoo (Déplopterus), the Spinetail (Synallaxis), and the Honey Creeper (Coereba). The interior of the eastern part of the island is a heavily wooded mountain, with its summit 3240 feet above the sea level, and always hidden by clouds. Here occur such forms as the Yellow-billed Thrush (Platycichla), Manakin (Chiroxiphia), Guan (Ortalis) and Parrot (Amazona). Some birds, such as the Creeper (Dendroplex) and the Antshrike (Thamnophilus), occur everywhere. The three westernmost islands, namely Aruba, Curagao, and Bonaire, have a small proportion of purely West Indian forms among their birds, in addition to the larger continental element (Hartert, 1893, 308). Margarita, however, which is much nearer the Lesser Antilles, has, according to Richmond and Robinson (1895, 653), an avifauna wholly derived from Venezuela, 27 kilo- meters distant. The islands closest to the coast have the largest variety of water birds. 122 OCEANIC BIRDS OF SOUTH AMERICA Thus the islets and reefs of Aruba, the driest of the three Dutch islands off Venezuela, harbor such species as Larus atricilla, Phaetusa simplex, Thalasseus maximus, Thalasseus eurygnatha, Sterna hirundo, Sterna dougalli, Sterna albifrons antillarum, Sterna fuscata, Sterna anaethetus, Phalacrocorax olivaceus, Pelecanus occidentalis, Fregata magnificens, and Phoenicopterus ruber. Margarita shares many of these, and has also the skimmer (Rynchops nigra). It will be noted that practically all of the above are species that also fre- quent the bights, beaches, coves, and lagoons of the continental shore. No one of them, except the Sooty Tern, is a cosmopolitan, blue-water type of bird. Most of them are of the American tropical and sub-tropical coastal assemblage, to which an island is preferable to a continental breeding and feeding ground solely because of its relative safety. Doubtless flocks of flamingoes, for example, were once as familiar on the Caribbean flats of Venezuela as they have since been on the ‘'salt-pans’’ of Bonaire CHartert, 1893, 335). Certainly pelicans prefer to nest as close to the mainland as possible. One of their largest colonies within the region is among the mangroves of Tortuga Island, which lies on the broad shallows of the coastal shelf, while the distant outlying groups, such as Islas de Aves and Los Hermanos, are without pelicans (Cory, 1909, 252). It is at such islets as the last named, together with the Roques cluster, that we find the stronghold of the oceanic birds par excellence. They are mostly small, scattered cays, though some are relatively high, and some are of sufficient size to enclose lagoons around which thrives a variety of low, halophytic vegetation. Certain of the islets at which boobies nest in greatest numbers, such as Orquilla of the Hermanos, are covered with granite boulders inter- spersed with dense growths of tall cacti which render human penetration ex- ceedingly difficult. Lowe (1909, 312) thus describes the main island of the Testigos, which lie on the line between Margarita and Grenada: Testigo Grande is of greater extent than is given in the sailing directions. We put its length at from three to three and a half miles, but its shape is very irregular, and in the short time at our disposal we had no opportunity of taking more accurate measurements. Its greatest height is (roughly) 400 to 500 feet. Scattered round it are other smaller islands or rocky fallarones [= farallones]. The group is periodically inhabited by fishermen, who go there from January to April, but we were informed that one family now lived there all the year. Testigo Grande is thickly wooded and covered with very varied vegetation, in spite of the waterless condition in which we found it. Many of the trees attain quite respectable proportions, and there seemed to me to be an unusual variety. Besides many that were unknown to me, I noticed Logwood, West-Indian ‘’Birch,’’ Guaiacum, Acacia, Tamarind, and Manchineel. Various kinds of Cactus grow on the lower slopes and on the smaller islets, and there is a great profusion of flowering bushes and even flowering plants. Wild cotton grows in abundance. Except for the few miserable huts erected by the fishermen, and a modest plot or so of cultivated ground in their immediate vicinity, these islands remain as they have existed through the ages. Geologically they consist of masses of a coarse-grained granite (hornblende), enormous fragments of which, much weathered, may be seen here and there amid the surrounding vegetation. On the weather side of one of the smaller islets this granite is overlaid by a layer of shaly rock. The general colour of the soil and exposed rocks in the lower parts of the island is deep ferruginous. I was much struck by the number of butterflies that we saw. Progress anywhere on the large island is laborious, owing to the thick bush and the excessive heat. CARIBBEAN: COAST 123 Localities of this sort are no doubt the least changed of the sea bird breeding stations in the Caribbean, and yet the ravages of goats and other introduced creatures may still have had an important unperceived effect. Most of these islands experience a concentrated, moderately heavy rainfall during the northern-hemisphere summer season, but they endure a long parched period afterwards. The breeding time of the majority of the sea birds seems to rise to a peak shortly after the end of the rainy season. Thus boobies of three species, man-o’-war birds, and tropic-birds are nesting in great numbers at Orquilla during January (Lowe, 1909, 324; 1911, 205). Noddies and Sooty Terns, which are evidently migratory, come to the island for the same purpose during February (Cory, 1909, 226). But the boobies, at least, breed during later months of the year as well, and it is likely that the nesting season for the population of any one species is protracted, if not practically continuous. At any rate, the offshore water of the Equatorial Current is of a character that suits the requirements of the boobies and certain other species throughout the year. It probably rarely or never falls below a temperature of 25° C., and it supports an abundance of the pelagic flying fish and cephalopods which are important, respectively, for the Blue-faced Booby (Sula dactylatra) and the Red-billed Tropic-bird (Phaéthon aethereus). These two species, as well as the Sooty Tern and the Noddy, cling mostly to waters beyond the coastal shelf. The Brown and the Red-footed Boobies, on the other hand, may in part follow the pelicans, man-o’-war birds, cormorants, and such species of terns as possess shore pro- clivities, toward yellower waters close to the continental beaches. Alexander von Humboldt (1852, 333) writes of ‘‘the bank abounding with fish, which surrounds the islands of Coche, Margarita, Sola, and Testigos; a bank of more than four hundred square leagues, extending east and west from Maniquarez to the Boca del Draco."’ Clark (1902, 260) states that both of the lesser boobies occur along the coast from Trinidad to La Guaira, and that the Brown species is as abundant in fishing grounds of the channel inside Margarita Island as it is around the islets farther from shore. At Aruba, Hartert (1893, 308) found Least, Roseate, and Bridled Terns all laying eggs during June, which would be after the close of the verano, or dry and windy season, and at the beginning of the rains. Petrels seem to be missing from the resident sea bird fauna of the Venezuelan islands. Audubon’s Shearwater nests as close to the region as Barbados and the Grenadines, but has yet to be recorded from the Venezuelan coast or islands. Wilson’s Petrels, however, enter this part of the Caribbean on their migrations, and Clark (1902, 260) reports a flock of a score or more, between Margarita and the mainland, on July 2. Belcher and Smooker (1934, 572; 1935, 279) have recently published a com- prehensive list of the birds of Trinidad and the neighboring Antillean island of Tobago, which forms part of the same British colony. The waterfowl which they list, with the status of each, are as follows: 124 OCEANIC BIRDS OF SOUTH AMERICA Puffinus lherminiert. Tobago, resident. Oceanodroma leucorhoa. Winter visitor. Phaéthon lepturus catesbyi. Tobago, resident. Pelecanus occidentalis. Resident, both islands. Sula sula. Tobago, resident. Sula leucogaster. Resident, Tobago and rocks off Trinidad. Phalacrocorax olivaceus. Visitor to Trinidad. Fregata magnificens rothschildi. Tobago, resident. Haematopus ostralegus palliatus. Tobago, resident or visitor. Larus atricilla. Resident, both islands. Phaetusa simplex. Trinidad, resident. Gelochelidon nilotica. Winter visitor. Sterna hirundo. Winter visitor. (Banded Massachusetts birds recovered in Trinidad.) Sterna dougalli. Winter visitor. (Banded Massachusetts birds recovered in Trinidad.) Sterna fuscata. Resident, both islands. Sterna albifrons antillarum. Trinidad, resident. Thalasseus maximus. Winter visitor. Thalasseus eurygnatha. Resident, probably at both islands. Anous stolidus. Resident, both islands. Anous minutus. Trinidad, possible resident. Rynchops nigra. Visitor to Trinidad, either from North America or Guiana. During the middle of the nineteenth century, Caribbean guano.from many of the lesser Leeward Islands assumed considerable commercial importance. Centinela, El] Roque, Los Monjes, and Los Testigos were among the hundred or more sites worked, and “‘myriads of waterfowl”’ are mentioned in the ships’ logs and other contemporary accounts. The product was a phosphatic guano of rocky type, with a laminated outer surface, often characterized by a high polish. It was therefore not similar to typical Peruvian nitrogenous guano, but rather resembled the ‘‘guano glass’’ found, along with a better grade, at the northernmost of the Peruvian islands, such as Lobos de Tierra, where rain occasionally falls (Taylor, 1857, 91). 2. THe ATLANTIC Coast, TRINIDAD TO THE SHOULDER OF BRAZIL The western side of the Gulf of Paria, and the coast that stretches southeast- ward toward the main mouth of the Orinoco, is low, heavily forested, and flooded during at least part of the year. The delta of the great river begins in the southern tip of the gulf, and to the eastward and southward of this point the shore has never been thoroughly explored nor more than approximately charted. The coastal waters are shallow and there are no towns or centers of trade to attract shipping toward the land between the few branches of the river that serve navigational needs. Even off the Boca Grande, the principal channel into the delta and the inland waters of the Orinoco, it is said that vessels have NORTHEASTERN ATLANTIC COAST 125 not infrequently grounded out of sight of land. The territory about the entrance is not only low, but is all so evenly covered with forest that it presents few distinctive marks to the eye of a visitor. Simpich (1931, 41), who took part in an air reconnaissance here, writes: Up yellow and brown estuaries of the dismal Orinoco Delta we flew, crossed intervening green jungle wastes, and returned to the muddy coastal tide flats by some other wandering stream in this swampy labyrinth. Except for two or three widely separated grass huts of savages, we saw no sign at all of human life. This is ‘The Land of a Single Tree’’ (the mangrove), as called by the Beebes (1910, 9) in a chapter from which the following passages are taken. One of the most astonishing things about the mangrove forest is the apparent diversity of its plant life. Until one actually comes within reach of trunk and leaves it is impossible to believe that all this forest is composed of a single species of plant. The foliage of some of the trees is light, of others, dark; here stands a clump of pale beech-like trunks, there a dark, rough-barked indi- vidual is seen. The manner of growth of the young and old trees is so different that a confusion of mingled trees, shrubs, and vines seems to confront one. But everywhere the mangrove reigns supreme. It is the only vegetable growth which can gain a footing in this world of salt water. In fact, it makes its own footing, entangling and holding mud and débris about its stems, and ever blindly reaching out dangling roots, like the legs of gigantic spiders. Far out on the tip of a lofty branch a mangrove seed will germinate, before it falls assuming the appearance of a loaded club from eight to fifteen inches in length. One day it lets go and drops like a plummet into the soft mud, where it sticks upright. Soon the tide rises, and if there is too strong a current the young plant is swept away, to perish far out at sea; but if it can maintain its hold, roots soon spring out, and the ideal of the mangrove is realized, the purpose for which all this interesting phenomenon is intended: the forest has gained a few yards, and mud and leaves will soon choke out the intervening water. The mangroves have still another method of gaining new territory. Aérial roots are thrown out from branches high in air, swinging downward and outward with a curve which sometimes wins three or four yards ahead. Like hawsers thrown from a vessel to a wharf these roots clutch at the mud beneath, but where the current runs swiftly they swing and dangle in vain, until they have grown so heavy that they touch bottom some distance downstream. . . . Throughout all this great region there is not a foot of solid ground. In one place we pushed a tall shoot some eight feet in height straight down through the mud, and it went out of sight. A man falling on this mud, out of reach of aid, would vanish as ina quick-sand. . . . More wonderful even than the coral polyps are these mangroves, for by this plant alone all this region has been rescued from the sea and built up into land. In future years, as the mud banks become higher and are fertilized by the everfalling leaves, other growths will appear, and finally the coast of the continent will be thus extended by many scores of miles of fertile soil. But the gain is specious; it is a case of robbing Peter to pay Paul. On the Atlantic coast of Venezuela and in parts of Guiana the land may, indeed, be extending its borders with wreckage from the interior of the continent, but nearer the Amazon it is probably being eaten away by the sea at least as fast as it is built up by fluvial detritus and plant growth. Agassiz (1868) regarded it as very remarkable that the Amazon, largest of rivers, has no delta—unlike any of the other streams we call great. By investigation he satisfied himself that most of the islands in the estuary are detached portions of the mainland, that in fact the sea at this point is breaking down the shore much faster than the river can build it up. There are, of course, a few small transiently growing 126 OCEANIC BIRDS OF SOUTH AMERICA islands of mangrove-covered alluvium, but in the main the process is one of disintegration. Von Humboldt (1852, 373) gives the following description of the plant and animal association of the mangrove thickets, and of the extension and eventual annihilation of the latter. Wherever mangroves grow on the sea-shore, the beach is covered with infinite numbers of molluscs and insects. These animals love shade and faint light, and they find themselves sheltered from the shock of the waves amid the scaffolding of thick and intertwining roots, which rises like lattice-work above the surface of the waters. Shell-fish cling to this lattice; crabs nestle in the hollow trunks; and the seaweeds, drifted to the coast by the winds and tides, remain suspended on the branches which incline towards the earth. Thus, maritime forests, by the accumulation of a slimy mud between the roots of the trees, increase the extent of land. But whilst these forests gain on the sea, they do not enlarge their own dimensions; on the contrary, their progress is the cause of their destruction. Mangroves, and other plants with which they live constantly in society, perish in proportion as the ground dries and they are no longer bathed with salt water. Their old trunks, covered with shells, and half-buried in the sand, denote, after the lapse of ages, the path they have followed in their migrations, and the limits of the land which they have wrested from the ocean. The Atlantic Ocean along this stretch of Amazonian, Guianan, and Vene- zuelan shore line receives a relatively enormous proportion of the earth's river water. This coast is the chief sluice-way of the wettest of continents. Heilprin (1906, 544) writes: Lest the geographer becomes too thoroughly impressed with the notion that the great rivers of the globe are only those that serve as trade-carriers or have played a familiar part in childhood’s study of geography, it may be noted that in this distant forest-land are rolling waters that would shame the rivers of Europe—that would put to a long test even the ‘Father of American Waters.” The great sweep of the Corentyn and Essequibo, rivers that 60 miles or more above their estuarine mouths measure from three to four miles in width, and which for yet a further hundred miles may still carry two miles, is an object-lesson in geography which impresses with more than ordinary force. The outpourings of one river overlap and mingle with those of the next one to northwestward, as all tend to fall in with the general course of the equatorial Ocean current. Sometimes even the special source of river water may be told by its color far from the mouth through which it issued. According to Smith (1879, 1), Amazon water, discolored by suspended particles of yellow clay, is determinable as greenish bands, parallel with the coast but from 300 to 500 kilometers away from it, and over deep soundings, as far as latitude 10° N. Farther south, near the equator, and still a hundred miles from land, the sea is much more strongly tinged; in April and May, indeed, it has nearly the clay-yellow hue of the Amazons itself, and furious currents struggle with each other until the surface boils and seethes as below a cataract. The flood of turbid waters, after this first battle with the ocean, gives way before the yet stronger equatorial current; its flank is turned, and it sweeps away northward, staining the sea with the blood of its defeat, littering it with débris, madly rushing into the heart of the enemy's country, until its last forces are exhausted and it sinks to annihilation, six hundred miles from the field of battle (Smith, 1879, 2). The experience of Pinzon, who reported a ‘fresh-water sea’’ off the northeast coast of South America after filling his casks out of sight of land, during the NORTHEASTERN ATLANTIC COAST 127 Amazon flood season of March, 1500, has been repeated many times. It is prob- able that much of the continental shelf, which is widest off the mouth of the Amazon, gradually narrowing toward Trinidad, has been built up of the count- less cubic kilometers of silt carried into the ocean by the Amazon, the Orinoco, and many lesser but still mighty streams. The action of the long-shore current is recognizable in old beachlines and spits, which are now obstacles to the direct flow of fresh water toward the sea, so that the smaller rivers tend to turn westward in the lower parts of their courses and to run for some distance parallel with the shore line. Off the mouth of the Essequibo, the pelagic water of the equatorial current, as distinguished from the mixed and turbid littoral water, has been detected within 40 kilometers of shore, in a depth of as little as 18 meters. Along most parts of the coast, however, the landward edge lies farther out. The offshore limits of the current are reputed to be in the neighborhood of 400 kilometers from land, and the axis of greatest velocity at a little less than half that distance. To return to the Orinoco delta, Boca Grande, the main mouth of the river, lying between Cangrejo Island and Barima Point, is about 16 kilometers wide. The approach is obstructed by an extensive bar which projects so far seaward that only in clear weather are the trees on the eastern side of the entrance visible from deep water. In windy weather there is a heavy and dangerous sea on the bar, which grows and sloughs away in such a manner that its depth is little affected by the height of the river, notwithstanding the great seasonal ampli- tude of the latter. The front of the Orinoco delta extends for about 260 kilometers along the coasts of the open Atlantic and the Gulf of Paria, and through numerous mouths, great and small, the river waters escape to the sea. Only the routes of the Boca Grande, the narrow Pedernales, and the Macareo are of commercial importance. Through the last named a steamship company operates its vessels on the most direct course between Port of Spain, Trinidad, and the inland city of Ciudad Bolivar, on the Orinoco. This route greatly reduces the distance, eliminates a voyage through the open sea, and so permits the use of river steamers and avoids the strong currents and troublesome waves of the Boca Grande bar. The seven or more other mouths are mostly little known. Several of them join at Barran- cas, about 180 kilometers up the main stream. After our consideration of such sister rivers as the Atrato and the Magdalena, it is hardly necessary to add that the delta of the Orinoco is not the home of birds that can be called marine. Petrels and boobies cleave to clean and saline water. Even Brown Pelicans and man-o’-war birds show no penchant for the shores of regions in which the sea water is excessively turbid. Occasional short files of pelicans may, indeed, be seen entering the estuaries of the dark delta channels, but the observer may be reasonably certain that they are wanderers and that their nearest breeding grounds are on islets of the bluer Caribbean. Only our adaptable old friend, the Bigiia Cormorant, seems as much at home in one place as the other. The avifauna, in fact, is made up of the characteristic assemblage of marsh 128 OCEANIC BIRDS OF SOUTH AMERICA birds which form fixed or transitory associations with the mangroves. Let us imagine the coming of evening in the Orinoco delta. Where were the myriads of water-fowl of which we had heard? We had seen nothing—not a single feather. But now the scene slowly changed. The tide was falling rapidly, swirling and eddying past the boat, and the roots of the mangroves began to protrude, their long stems shining black until the water dried from them. Mud-flats appeared, and suddenly, without warning, a living flame passed us—and we had seen our first Scarlet Ibis. . . . Before we could recover from our delight a flock of twenty followed, flying close together, with bills and feet scarlet like the plumage. They swerved from their path and alighted on the mud close to the mangroves, and began feeding at once. Then a trio of snowy-white Egrets with trailing plumes floated over- head; others appeared above the tops of trees; a host of tiny Sandpipers skimmed the surface of the water and scurried over the flats. Great Cocoi Herons swept majestically into view; Curlews and Plover assembled in myriads, lining the mud-flats at the water's edge, while here and there, like jets of flame against the mud, walked the vermilion Ibises. Terns with great yellow bills flew about the sloop, and Skimmers ploughed the surface of the tide in endless furrows. Macaws began to pass, shrieking as they flew, two and two together—and then night closed quickly over all. From the zenith the sun had looked down upon a stream as quiet as death; it sank upon a scene full of the animation of a myriad forms of life (Beebe, 1910, 6). By the end of March the waters of the Orinoco begin their vast annual rise, attaining greatest height in Juiy or August and gradually falling away again to a minimum in February or March. The ordinary range in water level amounts to as much as 18 meters. A range only slightly less is known even at Ciudad Bolivar, 367 kilometers upstream. With the flood, the current of the river increases, and the ocean surface off the Boca Grande is said to become quite fresh 30 kilometers from land. At the time of heaviest rains the northern parts of the delta, where the outlets run into the Gulf of Paria, are covered with water that has overflown from the channels, and the human inhabitants take to dwellings built in trees. The block of territory extending from the scarp just south of the River Orinoco to that just north of the lower Amazon is collectively known as Guiana. It is one of the geologically ancient plateaus of the continent. Historically it is divided into Venezuelan, British, Dutch, French, and Brazilian sections. The foreshore from near the Boca Grande of the Orinoco to the Oyapock River at the Brazilian border continues generally low, being formed of alluvial matter brought down by very numerous rivers. Toa considerable extent the land close to the sea is slightly below spring tides, which flood wide areas; at low water the shores of many estuaries appear as inaccessible banks, while at high water they are inundated. Certain settlements, such as Georgetown, are diked to meet this condition. The submarine slope of the muddy coast is so gradual that depths of 50 meters may sometimes be found almost as many kilometers from the land. Barely submerged mudbanks extend offshore in places as far as 15 kilometers. Such banks are constantly shifting under the influence of river and ocean cur- rents. Clumps of vegetation carried down during the rainy periods not infre- quently become established after stranding on shallow parts of the banks, with a resulting growth that alters the features and even the direction of the coast. The pilot books are filled with references to points of land in Guiana that can NORTHEASTERN ATLANTIC COAST 129 no longer be made out. Moreover, owing to the monotonous uniformity of the coast, every ephemeral feature, such as a clump of burned trees, is seized upon as a landmark. Schomburgk well described these shifting shallows nearly a century ago. The mudbanks, so dangerous to coastal navigation, he writes, are the nuclei of ever-nascent formations along the Guiana coast. Their magically rapid birth and disappearance has something mysterious about it. Where a few days before was a smooth surface of water, there abruptly appear soft mudbanks which, after a short time, may be as quickly washed away. Frequently, how- ever, they are pushed bodily against the shore where, out of reach of the general drift, they come to form part and parcel of it through the agency of mangroves and related plants. The whole coast line is thus subject to constant change, and many a settler who could formerly look from his windows out to sea now finds the view shut off by a forest. Further confirmation of the continual in- crease of foreshore is indicated by the successive tiers of coast vegetation which, according to their respective ages, rise one above the other inland like the seats of an amphitheatre. The alluvium cannot come in large part from Guiana, the streams of which are bright and translucent except in the relatively narrow coastlands. If the latter had been the source, they would have been all washed into the sea ages ago, for the littoral ocean is so chock full of mud and mire, along an area at least 240 kilometers in length and 16 out from shore, that in some places the very waves resemble soft thin ooze (Schomburgk, 1847, 1, 187). The pilot books make much of the “‘dark suspicious hues’’ of the coastal water of British Guiana, which is due to a combination of the silt, the shoal and muddy bottom, and the vegetal staining of the river increment. During a coastwise trip out of Georgetown, Beebe (1910, 134) noted the inhibiting effect of suspended matter upon wave movement: ‘The water along the coast,"’ he writes, ‘is very shallow and is so filled with sediment that even in a heavy gale the waves break but little.”’ Schomburgk (1847, 1, 252) has further described the egress of the muddy Essequibo through an estuary 20 kilometers in width. The grand spectacle of the ocean was increased by the sight of the mighty Essequibo. Although the mouth of this river was still several miles away, we could already see the powerful waves of the stream flow like a dirty ribbon out to the sea, without mixing with the waves of the ocean. Still more lively and impressive this became when at high tide the two immense water forces met in a wild fight and stirred up the surface to an incredible tumult—but the proud river proved its superiority and soon conquered the power of the tide; the high foamy waves and whirling calmed down, and the glorious ribbon of the victorious river extended before our eyes again, and was only lost to view far away at the horizon. And Heilprin (1906, 539) carries on, in the following pleasant introduction to the country, the references to the earthy cargo of the littoral waters which, indeed, few visitors to that part of South America have failed to note. Most travellers receive their first impressions of British Guiana on their arrival at the capital city, Georgetown (more commonly designated from the country in which it is situated, Demerara), but before this, and for many miles out to sea, he will have noted a characteristic of the land in 130 OCEANIC. BIRDS OF SOUTH AMERICA ~ the yellow-brown waters that lie outside, the sediment that rolls out in floods or is gathered in from the discharge waters of the Orinoco. The Amazon seems to have little part in this discolora- tion; for, if my information is correct, much of the off-shore of French Guiana and the land lying still farther to the eastward are bathed by a clear sea. Back of the fringe of muddy water the eye follows the low contour of a bordering grove of mangrove, and beyond it in some places, or re- placing it elsewhere, the glass may resolve a somewhat loftier growth of willow-like bushes, or even trees, the courida (Avicennia nitida). In the distance, less lofty than the occasional chimney that tells where the sugar-cane is being grown, or where it was cultivated until competition with the sugar-beet broke the back of the most important, albeit steadily falling, industry of the colony, are a few specimens of the royal and cabbage palms (Oreodoxa regia, Oreodoxa oleracea), and the land- scape is virtually complete. There is not a rise in the land that even remotely suggests a hill, and it is in faith only that the mind constructs the landscape of lofty mountains in the far interior. The muddy water of this coast is, in my opinion, the factor that limits the southeastward extension of the Brown lican’s distribution so abruptly in the neighborhood of Trinidad. On the west coast of South America pelicans have taken advantage of continuously favorable littoral conditions, and have pushed their specific range southward to central Chile. Much of the eastern coast of the continent appears to offer an even better habitat for pelicans, despite the fact that none are present. To southward of Cape Sao Roque, in particular, inlets like Bahia, the Bay of Rio de Janeiro, and the straits behind the shelter of islands such as Sao Sebastiao and Santa Catharina, have every appearance of places in which pelicans could find an abundance of food, clear and bland water, and a climate admirably suited to their welfare. Furthermore, the Abrolhos Islets would seem to offer breeding sites to suit every need. Rocas Reef and Fernando Noronha are doubtless too far from the continent to be equally available. In every other respect, however, they too would make ideal “pelican islands’’; neither should it be forgotten that this same species of bird has at some time in the past established itself upon the much more remote Galapagos. I would hazard the guess that if fifty adult pelicans from the West Indies were set free along the coast of southern Brazil, the birds would take to their new home as rabbits did to Australia, and would quickly found a flourishing colony. But the ever-present and insuperable barrier to their natural distribu- tion into this region is the mud and silt that fills so much of the coastal ocean between the Gulf of Paria and the mouth of the Amazon. In this turbid water there are either no schooling fish in numbers sufficient to support a population of pelicans, or else the water itself is so nearly opaque that the pelicans are unable to see and capture their prey. A critical reader may object to this that the boobies and the man-o’-war bird, which have succeeded in rounding the eastern tip of South America, might presumably have been thwarted by the same unfavorable conditions off the Guianas. The answer is that birds of both these groups have more pronounced pelagic inclinations than pelicans. Perhaps it would be more correct to say that the boobies and man-o’-war birds have less pronounced inhibitions against leaving the immediate neighborhood of the coast; probably, indeed, a man-o’- war bird would voluntarily fly 50 kilometers from shore more readily than a NORTHEASTERN ATLANTIC COAST 131 pelican would fly 10. It is likely that for such reasons the West Indian forms of Fregata and Sula have crossed or circumvented the barrier of silty waters, to reappear as nesting birds at Fernando Noronha and in the blue-water paradise along the southern tropical coast of Brazil. For a number of reasons the man-o’-war bird would be much more at home than the pelican on the coast of Guiana. Among these are the following: (1) Although it roosts by night in the mangrove forest, it is ready and willing to fly a considerable distance out to sea in search of favorable fishing waters; (2) it is not wholly dependent upon prey captured by itself, but feeds in part upon food which it forces other birds to disgorge; (3) it does not dive for its food and is therefore not limited to transparent water—the fish it captures for itself are mainly forms which come to the surface, or which even leave the water to make short glides in the air. Little seems to be known about breeding stations of the man-o'-war bird between the Caribbean and Cape Sao Roque, but it is probable that the species nests, or formerly did so, on some of the islets off the coast of French Guiana where, it is said, the water over the continental shelf is prevailingly more clear than that to westward. Boobies and tropic-birds are all scarce, if not quite unknown, along this section of the South American coast. Tropic-birds would, of course, entirely avoid muddy, or even discolored, water. The same is true, though perhaps to a lesser extent, of the boobies. Moreover, only one species of the tropical Atlantic group of boobies is known to roost consistently in trees. The tropic- birds and boobies are, however, sufficiently pelagic to have worked around the outer edge of the turbid coastal zone, so that they reappear farther southward. Finally, there is another consideration to be mentioned which has a bearing upon the foregoing hypothesis, without invalidating its application to the pelican. The Atlantic members of Fregata, Sula, and Phaéthon are all, to a greater or less degree, birds of cosmopalitan type, and there is no reason to believe that they have originated in the West Indian region, thence to spread into other parts of the world. It is rather more likely, on the contrary, that they origi- nally reached the Caribbean by invasion in the direction towards which the converging trade winds blow. The Brown Pelican, on the other hand, either originated in the Caribbean-West Indian district or else it has been extirpated elsewhere. In either case the opportunities for it to spread southeastward, toward suitable waters on the Atlantic coast of Brazil, are greatly lessened by the same natural forces which once made ‘rounding Cape Sao Roque’’ a difficult feat for sailing vessels, namely, the constant opposition of winds and a strong ocean current. When we add to these obstacles the presence of a band of muddy water, and the Pelican’s congenital disinclination to pass out of sight ot the coast, the limits of its present range in this direction are not hard to compre- hend. Brown Pelicans have, of course, been reported from the coast of the Guianas and even trom inland waters of the Amazon. All such records, how- ever, appear to be based upon wandering individual birds. The species is not mentioned in most lists of residents of the region, and Schomburgk states that it does not nest anywhere southeastward of the Venezuelan coast. 132 OCEANIC BIRDS OF SOUTH AMERICA Between the mouths of the Orinoco and the Essequibo the low coastal region is filled with large lagoons and marshes, the resident Indians construct- ing their dwellings on posts well above the ground, and travelling about largely in ‘‘woodskins,"’ as the Guianan canoes are called. On the open shelving sea beaches a heavy though sluggish surf is usually in force, especially during flood-tide, which augments the current and the run of the waves. In places, as at the Waini and Moruka beaches, there are conspicuous beds of shells, which are a commercial source of lime, but owing to the breadth of the shallow flats only small vessels can come inshore close enough to load. Shells of mol- lusks are more or less mixed with the littoral mud all along the coast of Guiana and have been made the basis, by Schomburgk and others, of speculations regarding the source of the growing foreshore. The coastland bird life of the common border of British and Dutch Guiana is thus described by Young (1929, 751) who, it is interesting to observe, men- tions neither the pelican nor the man-o'-war bird in his account: The littoral and riparian area embraces the bog-land of the Corentyne Coast, the mangrove swamp-jungle through which the rivers flow and which also occurs at intervals on the shore, and the wide spaces of mud exposed at low tide and bordered on the landward side with arrow- grass and shrubby trees. The swamp-jungle consists typically of the two mangroves (Lagunularia and Rhizophora), the courida (Avicennia nitida), which lends its name to this type of jungle, known to the creole as ‘‘courida bush"’ and the impenetrable ‘‘bunduri pimpler’’ (Drepanocarpus lunatus), the name ‘‘pimpler’’ being applied to any plant with thorns. Where this bush has been felled by man the exposed mud becomes covered by the giant swamp-fern (Achrosticum aurem), which grows in clumps five or six feet high. The only species of bird peculiar to this swamp-jungle is the Hoatzin (Opisthocomus hoazin), which spends its whole life within a comparatively smallarea. . . . The Corentyne Coast bogs constitute the hunting-ground of the Brazilian Cormorant (Phala- crocorax vigua), and here, as well as on the open flats, the migratory Waders disport themselves— Spotted Sandpipers (Actitis macularia) in clouds, and others less abundant, such as the Solitary Sandpiper (Tringa solitaria) and Ringed Plover (Charadrius wilsoni). Occasionally companies of White-faced Tree-Duck (Dendrocygna viduata) paddle about on the water’s edge, and groups of the beautiful pink-legged Stilts (Himantopus mexicanus) may also be met with. This is a fascinating region, and the enjoyment of watching the most varied assemblages of shore- and water-fowl is often the reward of anyone who braves the prospect of struggling up to the knees in ooze, stabbed mercilessly by arrow-grass and subjected to the undivided attention of clouds of mosquitoes and sand-flies. Quelch (1896, 258) likewise refers to the great flights of North American migrants in the marshlands, especially during former times when the Golden Plovers (Pluvialis dominica) began to arrive in force during August, increasing thereafter until October, and then diminishing until the last stragglers departed southward in November. During the dry season, according to Quelch, the mative waders are mostly confined to the stream courses or to isolated ponds, but after the wet season begins, the ““Negrocop”’ or Jabiru, the Spoonbill, and other marsh birds disperse over the flooded savannahs of British Guiana, where they feed upon aquatic creatures spread abroad by the floods or upon insects and reptiles driven out of the undergrowth. The climate of British Guiana approaches the sub-tropical, despite the proximity of the equator, the sun's heat being to a large extent tempered by NORTHEASTERN ATLANTIC COAST 133 the fresh northeast sea breezes, which blow practically daily for the greater part of the year. There are two wet seasons annually, a longer one from April to July and a shorter one in December and January. At these times a southwest or land breeze is often experienced, bringing with it the heaviest falls of rain. In Dutch Guiana, according to the Penards (1908), the long rainy season begins slightly later—in May or June. The savannahs become bogs and the creeks swell into terrifying torrents. Toward August, the rain slackens and sunshine quickly follows the showers. This is the time of year when young ibises, herons, flamingoes, and other native waders spread in numbers along the coast. By September relatively fair weather prevails; the ground dries and cracks; temperatures are high; easterly or northeasterly winds blow during the afternoons. This is the short dry season, to be succeeded about the middle of November by a short wet season, during which there are heavy morning fogs and comparatively low temperatures. The annual precipitation increases toward the eastern end of the region, reaching a maximum in French Guiana. There the so-called dry seasons scarcely deserve the name, judging by rainfall records, except that September and October have little rain at the coast, while August and November are at least drier than the remaining months of the year. To this extent we see an approach toward the rainfall pattern of the northeastern tip of Brazil and the island of Fernando Noronha. Along the coasts of British and Dutch Guiana there is a dearth of outlying islands as well as of eminences on the mainland. In French Guiana the same condition holds as far eastward as the Iracoubo River. At this point, however, the hills of the same name become visible from the ocean, and farther eastward isolated peaks or low ranges are much in evidence. Likewise, small groups of islands begin to appear, most of them being similar in formation to the fringing mud flats of the coast. Cayenne Island, however, rises to 255 meters above sea level. It is lined on one side with bold rocky coves and steep islets, presenting a strong contrast with the ports of the neighboring Dutch and British colonies. This prominence is essentially a mainland spur, but offshore in the same region are the Salut Group, with its notorious Devil's Island, the Battures, the Conné- tables, the Pére and Mére, and |’Enfant Perdu, the last so called because it is a long way to leeward of its parents and seems as if it had strayed hopelessly from home (Waterton, 1839, 94). Bellin, in 1743, spoke of these islets as “sterile rocks,’’ but recent photographs (Simpich, 1931, 45, 46) show that several of them now have good growths of palms and other vegetation. One of the Connétable Islands is said to be worked for building stone and guano. It would be interesting to learn the source and nature of guano which has ac- cumulated in so humid a region. As to the abundance of birds, however, there seems to be no doubt, for Barrere, another French chronicler who, like Bellin, published in the year 1743, says that the Great Connétable is known also as Bird Island, and that frigate-birds and boobies occupy it. It has remained a. sanctuary for sea fowl, he continues, because the strong tides and currents make it hazardous for sailing craft to reach it. Waterton(1839, 95) adds the following: 134 OCEANIC BIRDS OF SOUTH AMERICA Considerably to windward of Cayenne, and about twelve leagues from land, stands a stately and towering rock, called the Constable. As nothing grows on it to tempt greedy and aspiring man to claim it as his own, the sea-fow] rest and raise their offspring there. The bird called the frigate is ever soaring round its rugged summit. Hither the phaeton bends his rapid flight, and flocks of rosy flamingos here defy the fowler’s cunning. If boobies and tropic-birds are, indeed, present at the offshore islands of French Guiana, it would tend to confirm Heilprin’s statement, quoted above, that the littoral ocean is here clearer than the waters either to eastward or to westward. Barrere remarks, furthermore, that this coast is particularly rich in fish, and Waterton that it offers ‘‘by far the sublimest scenery on the seacoast from the Amazons to the Oroonoquo.”’ The Oyapock River, at the Brazilian boundary, has a funnel-shaped estuary about 20 kilometers wide. Beyond the more or less swampy lowlands to west- ward of the entrance, the terrain rises into picturesque hills and mountains that appear from a distance like islands. The country between the French border and the Araguary River, at the mouth of the Amazon, is the sea front of Brazilian Guiana. Behind the uniform-look- ing coast, with nothing to indicate the position of most of the rivers, there are many large lagoons lying between shore and highlands. The whole region is more or less inundated during part of each year, and is covered with mangroves which can be seen at a distance of not over 15 to 20 kilometers. Being exposed to the scouring action of waters discharged by the Amazon and other rivers, the outline of the coast undergoes continual change. Soft mud flats formed by the sediment of the rivers extend off in many places considerable distances. These the mangroves invade rapidly during the dry season, often to be torn up again by the violent action of currents when the period of freshets ensues. Moreover, the high tides, advancing up the streams far into the country, smear the banks with an enamel of slimy, fine-grained mud. Goeldi (1896, 97) describes a boat journey during the month of October from Para to the River Counany, which lies about midway between French Guiana and the Amazon. The coastland, he writes, looked from sea like a narrow blue-green ribbon, surprising in the apparent regularity of its breadth. The ocean water was entirely of a muddy brown hue, even though the date was half a year after the Amazon flood season. When he reached the mouth of the Counany, the tide was out, and even his very small steamer was unable to enter until it had risen, carrying with it toward shore many reddish brown jellyfish. Even with the flood-tide, progress was difficult because of bars and uprooted trees, all hidden by water which the author describes as little more than fluid mud. The forest on either bank was of the typical Guianan sort— siriaba, with occasional courida trees, tall palms of two species, clumps of bamboo, and many lianas. After an hour and a half the coastal or navigable stretch of the river had been passed, and Goeldi continued his voyage by means of a dugout. Sub- sequently he ascended a neighboring stream, the Amapé, where conditions were equally muddy, dreary, and devastated, with piles and tangles of tree trunks NORTHEASTERN ATLANTIC COAST 135 along the banks, wreckage of the terrible ‘‘pororéca’’ or bore. For a proper conception of the “‘miles of giant forest trees, uprooted and scattered like matches’’ by rivers of this coast, the reader should refer to a photograph from the air, published by Simpich (1931, 47). In this part of the coast, and southeastward, the tidal bore is perceptible in the estuaries of most of the shallow rivers. In the Calsoene River spring tides rise 7 meters, and at Maraca Island 9 meters. The Calsoene is incidentally characterized by a curious tidal phenomenon, called the ‘“‘doucin’’ because the water in the estuary becomes entirely fresh. After heavy rains there is no longer a flood current; even though the tidal level rises and falls as usual, a vessel at anchor remains constantly swung to the ebb. The mouths of the Amazon extend along about 290 kilometers a ocean front. Despite the amount of silt carried by the great river, far larger in the volume of its water than any other, it has built up no projecting delta like those of the Mississippi and the Nile. Its deposits appear to be carried away con- tinuously by the action of currents flowing in a northwesterly direction along the coast and attaining, offshore, a maximum speed of 5 to 6 kilometers per hour. Much sediment is evidently dropped, at least temporarily, at the north- western border of the estuary for here, off Cape North, a vast shoal stretches seaward, the 10-meter contour being in some places as far as 80 kilometers from shore. A certain mixing of river and ocean waters also takes place on the south- easterly or up-current side. Agassiz (1868) noted that when he approached the mouth of the Amazon from the south the first indications of the river were yellowish patches staining the sea here and there. Presently the patches were replaced by broad streaks, after which the fresh waters began rapidly to super- sede the salt, while the steamer was still several hours’ journey to southeast- ward of the Para estuary. The dark water of the Amazon discolors the ocean for 80 kilometers or more from its mouth. The annual inundations in the lower part of the basin occur chiefly between January and May, the rise of the river amounting in some places to as much as 15 meters. Tidal effects are extremely varied and confusing in the complicated waterways of the estuary, the powerful bore sometimes occurring in the shallow waters of the western mouths and elsewhere, at or near spring tides. In maritime instructions the bore has been characterized as ‘‘a head of water five to twelve feet in height, with a breaking front, advancing at speeds which range from ten to fifteen knots.’’ Under such circumstances, the entire tidal rise along the shores of the northwestern channel is said to occur within the short space of ten minutes. Smith's (1879, 12) description of the Amazon bore is somewhat less spectac- ular and is doubtless more factual. Near the mouth this wave is very apparent. The tide is forced, so to speak, into a funnel, over shoals and against the descending current; it rises in a great solid mass three or four feet high, uprooting trees along the banks and breaking canoes that may happen to be in shallow water. This is the celebrated pororéca, a phenomenon which is best seen on the northern side of the river, 136 OCEANIC BIRDS OF SOUTH AMERICA and during the spring tides. Travellers have had much to say of the pororéca, and some of them, no doubt, have multiplied it in their fancy. However, the tidal wave is really formidable, and much dreaded by the canoemen, who keep in mid-channel to avoid its force. It is largely because of the presence of the bore that the western or main entrance to the Amazon is little used for commerce, the usual maritime route being by way of the channel leading toward Para. Alfred Russel Wallace (1853), however, refers to formidable manifestations of the ‘‘piroréco,”’ as he spells it, even in this branch of the river. On one occasion during his youthful field work in Brazil, he had gone ashore at a sugar estate to wait for the incom- ing tide, when the agent told him to take the boat out into the stream in order to avoid the pororéca. Just as the party was expecting the tide to turn, a great wave came rushing along and broke at the place where the canoe had been previously moored. When the wave passed, the water was as quiet as before but flowed up with great rapidity. In winter when the spring tides are highest, the bore breaks with terrific force, according to Wallace, often sinking boats or dashing them to pieces if they are left in shallow water. His explanation is that when a body of tidal water in rapid motion is of sufficient depth to come in contact with a shoal, a great wave is formed which continues for a long dis- tance beyond. For a day’s journey or more up the great river, the onrushing tidal wall causes the waterfowl, and the marsh birds on the banks, to take precipitant flight. Stretching across the main mouth of the Amazon is a chain of islands, of which the two largest are Caviana and Mexiana. The latter is the most southern and seaward of the group, and is separated by the Canal do Sul from the much more extensive Ilha Maraj6é, which is really to be considered as part of the riparian mainland, despite its separation on the western side by a complicated web of creeks. Mexiana has been visited, and reported upon, by a number of naturalists, including Wallace in 1853, Hagmann in 1908, and Miller in 1914. Since it is inhabited by all of the water birds of the coastal region, as well as by a rich variety of land birds, we may pause to picture it as a typical example of Atlantic equatorial lowland fronting on the ocean. Mexiana is oval, about 50 by 30 kilometers in dimension, and is bisected by the equator. Despite the fact that its northern and eastern sides face the open Atlantic, it is surrounded only by fresh water; neither saline nor brackish water is noticeable at any point, even when the Amazon attains its lowest level dur- ing drought periods. The greatest elevation is only a meter or two above mean sea level, and the very highest points comprise little mounds called “‘tesos,”’ which are covered with tall growth and which thus dot parts of the interior grassy campo with ‘‘islands’’ of forest, in which the brilliant yellow blossoms of the knotty carobal trees come into bloom during August. The general level of the island is highest near the shores and falls away to- ward the middle, this being due to the fact that during flood seasons the larger part of the overflowing load of sediment is deposited around the borders. Because of this same annual enrichment the principal belt of marshy forest, containing palms, latex-producers, and other lofty trees, as well as an abun- NORTHEASTERN ATLANTIC COAST 137 dance of bamboos and lianas, is likewise marginal, the whole center of the island being grassy or shrubby campo. In a few places this breaks through the forest girdle to extend as open glades down to the muddy or sandy beaches, which at high tide are mostly submerged. Numerous small meandering creeks or ‘‘igarapés,’’ which are without sources on the island, receiving all their water from the Amazon, penetrate inland from various parts of the coast. These have dense gallery-forest along their lower reaches. They are all subject to tidal rise and fall, and many of them lead to fair-sized lagoons in the interior sinks of the island. Toward the end of March such bodies of water greatly enlarge with the floods. Thereafter some of them dry up, or remain only as swamps or ‘“‘mondongos”’ during the relatively long dry period. As a rule no rain falls from about the middle of August until early January. Mexiana is all thoroughgoing alluvial land, mostly clay and sand, with no pebbles or other stone. The variety in the soils and the forest vegetation in different parts of the terrain led Miiller (1914, 35) to the opinion that the island has originated through the consolidation of several smaller estuarine islands, and that the principal igarapés are the last remnants of former dividing channels. On the whole, Mexiana gives a much more tropical impression than Marajo, although the two are separated by only 8 to 9 kilometers of river. Mexiana is locally celebrated for the abundance of its alligators, ounces, and birds. Among its forest trees live many cuckoos, anis, paroquets, hummingbirds, woodpeckers, tanagers, cotingas, and birds of prey. There is also one species of toucan (Rham- phastos toco), and the Black Vulture is common. Flocks of Muscovy Ducks fly with a noisy rush of wings toward the lagoons as long as any water remains in them. Cormorants plunge in the deeper, surrounding channels. Sun-bitterns and Spur-winged Plovers feed in the igarapés, especially along their jungle- bordered stretches, and a host of marsh birds, such as Limpkins in search of snails, Spoonbills, Scarlet Ibises, Jabirus, Wood Ibises, and several species of herons appear in great numbers with the coming of the showers, about New Year, and remain throughout the wet season until July, when many of them retire to more permanent watery fastnesses of the mainland. In the coves and canals around Mexiana one can usually see flocks of river terns of several species, as well as Skimmers (Rynchops nigra cinerascens). The Gull-billed Tern (Gelocheli- don nilotica) is also said to nest at the island. There is thus no resident sea bird fauna. On the offshore side one might see an occasional man-o’-war bird in pursuit of a skittering ““balao’’ (Hemirhamphus or a related fish), but the Mexiana assemblage as a whole is rather of the sort with which we have already become acquainted in delta and marshland regions along the Caribbean. In short, with the notable absence of the Brown Pelican, it is the standard avifauna of humid tropical coastlands in South America, not very different from what we might find along waterways of the Gulf of Guaya- quil, on the opposite side of the continent and close to the same latitude. That the waters off the mouth of the Amazon are, nevertheless, a favorite region of sea bird concentration within the tropical Atlantic is clearly indicated by observations of the ‘Meteor’ Expedition referred to above (p. 89). Hent- 138 OCEANIC BIRDS OF SOUTH AMERICA schel’s chart (1933, 117) shows seven stations, between the mouth of the Rio do Para and a point nearly a thousand kilometers to northeastward, in which the oceanic bird count is relatively very high, as compared with neighboring parts of the tropical ocean. Birds were less common near the land than well offshore; at five of the pelagic stations between 10 and 19 individuals per watch period were observed. Hentschel does not record the species of birds observed within this oceanic tongue, which is enclosed by the parallels of 0°-6° N., and the meridians of 40°-48° W. In the same neighborhood during April, however, I have seen numerous storm petrels (both Oceanites and Oceanodroma), Greater Shearwaters (Puffinus gravis), unidentified terns (perhaps Sterna paradisaea), boobies, tropic-birds, and a skua (Catharacta). Examples of the petrels were collected. The skua flew between the masts of the brig ‘Daisy,’ in latitude 5° 54’ N., longitude 40° 30’ W., on April 30, 1913. Simmons (1927, 75) likewise reports skuas between the equator and the West Indies during mid-May. Ocean waters receiving such a vast discharge as that of the Amazon should, of course, offer favorable conditions for a rich neritic and high-seas life, since they possess at least three of the environmental advantages which Hentschel has listed at the conclusion of his report on oceanic biology (1933, 164). In this region, for example, there are extensive areas of reduced salinity and of enormous terrigenous additions to the ocean. There are also local zones of sinking and of upwelling, caused by differences in temperature and density. And, finally, there is a very active mixing of waters, due to the penetration of the ocean by a sustained flow from the greatest of rivers. All such circum- stances, as pointed out heretofore, tend to stimulate a succession of life in the sea. That offshore currents sometimes show pronounced temperature differences even in the equatorial Atlantic is shown by the following observation made by Ball (1887, 356) on the evening of July 30, in a latitude a little to southward of the mouth of the Amazon. With the thermometer standing about 82° [F.], the passengers naturally preferred the upper deck to the close air of the saloon, and were resting in their ship-chairs between nine and ten p.M., when suddenly there came an outburst of coughing and sneezing, followed by demands for muffling of every kind. There was no sensible movement in the air, but I found that the thermometer had fallen to 79° Fahr., and there was a feeling of chilliness which was not easily explained by that slight fall of temperature. The mystery was explained on consulting the chief officer, who throughout the voyage paid much attention to the temperature of the sea. Since leaving Pernambuco, the thermometer in a bucket brought up from the surface had varied only between 82° and 83°. On this evening we had abruptly encountered a relatively cold current, with a temperature somewhat below 76°, and the effect of being surrounded by a body of cool water when the skin was in the condition usual in the tropics was felt by nearly all the passengers. The estuary of the Para, which is in reality the mouth of the Tocantins rather than of the Amazon, is so wide (about 55 kilometers) that from a mid- position the wooded banks on either side are out of sight. Mud and mangroves, flats and narrow waterways mark all the shores, and the river craft produce a certain resemblance to a Chinese stream. Vast extents of territory, such as part of the great pasture land on Marajé Island, are flooded during the rainy season NORTHEASTERN ATLANTIC COAST 139 which, as explained in the section on the Amazonian climatic district, is be- tween December and April. From here eastward toward Cape Sio Roque, how- ever, the peak of the wet period comes later and later in the year until, from February at Belem, it reaches June at the shoulder of the continent. All of the northerly coastland of Brazil is low, except in the vicinity of Ceara. East of the Amazon the broad continental shelf, studded with green islands close to shore and with small shallows farther out, narrows only gradu- ally toward the east. Behind the beaches are many sandhills visible from the coastwise steamship routes. They present a similar and monotonous aspect, although occasionally interspersed with small reddish cliffs and with clumps of mangroves on the western banks of the rivers. Because of wind and current trends, it is characteristic that the western or leeward side of the mouths of the streams should be covered with vegetation while the eastern sides are sandy and barren. Sometimes the lower sand dunes look like breakers; again, as at Salinas Falsas Bay, they rise in such a manner as to resemble the white sails of schooners. Still farther east are stretches of coast on which the sandy areas look like gigantic pieces of white linen laid out to dry. Northwest of Maranhio, for example, are the islands known as Lengdes Grandes and Peque- nos, or the Big and Little Sheets, names likewise repeated on the mainland farther eastward. Behind the wave-built ridge of the actual beach, which in many places serves as the chief coastal highway, lies much broad, green-forested lowland, with countless meandering rivers. The mouths of the latter are for the most part so obstructed by bars that truly navigable streams are few and far between. Most of the outlets, moreover, seem to be progressively shoaling so that ports like Aracaty, formerly frequented by foreign commerce, are today accessible only to small coasters. There are few safe anchorages for vessels along the open seacoast, and even boat-landings can be made only where spits or reefs extend out from shore, thus forming on their western sides a shelter from the constant swell. Vessels have been detained inside the bar of the River Jaguaribe for months, for want of sufficient water to float them across. A large part of this coast is imperfectly surveyed, but at any rate changes are so rapid that constant revision of charts would be necessary. Between Sao Luiz do Maranhao and Cape Sio Roque the only good harbor is Tutoya Bay. Studies of this part of Brazil—its physiography, vegetation, and bird life— have been made by Snethlage (1917, 41), Reiser (1926, 107), and Hellmayr (1929, 241). In northern Para and Maranhio most of the hinterland is still covered with majestic primeval forest, which verges into ‘‘campo’’ toward the east. Close to the seashore and along the larger rivers, wooded swamps or “igapds’’ prevail in even the generally open districts of these states. According to Hellmayr, the coastal forest can be traced as far east as Miritiba, but some- where between here and the Rio Parnahyba the woodland gives way to an open or at least cleared terrain. Gallery forest naturally continues to extend inland along the rivers well into the otherwise treeless country. Hellmayr lists, as characteristic of rivers or other waterways along the 140 OCEANIC BIRDS OF SOUTH AMERICA coast, the Pied-billed Grebe, Bigiia Cormorant, Anhinga, Laughing Gull (doubtless a migrant from the north), the large and small river Terns (Phaetusa Simplex simplex and Sterna superciliaris, respectively), and the southern race of the Black Skimmer (Rynchops nigra intercedens), of which the type locality is Sao Paulo. Inhabiting the beaches during northern-hemisphere winter months are such migrant shore birds as the Hudsonian Curlew, Spotted Sandpiper, Sanderling, Least and Buff-breasted Sandpipers, Robin Snipe, Semipalmated Sandpiper, Dowitcher, Wilson’s Plover, Black-bellied Plover, and Ruddy Turnstone. For some distance to eastward of the Amazon the coastal waters retain a yellowish tint, due to the sediment of many rivers, but toward Ceara they become progressively clear and acquire a greenish aquamarine hue which is famous. The change is due both to greater clarity and to the replacement of dark and silty bottoms by sand. It becomes most pronounced near the point where the relief increases, and where the streams consequently wind through less marshland on their course toward the ocean. It is a short distance east of the delta of the River Parnahyba that the first highlands become visible from the sea and shore—the northern end of the Ibiapaba Mountains, which loom a thousand meters above a sandy foreground. Lesser hills along the coast over- look a great expanse of plains and swamps and a cluster of peaks reaches the shore at Fortaleza. A heavy sea and a powerful westward-setting current prevail almost con- stantly along this part of the coast. A frequent haze, and the embarrassing uniformity in the appearance of the Jand, add to the trials of navigation. The heart of this district, in Rio Grande do Norte, is the Brazilian ‘Job of the North,”’ discussed in the meteorological section. Southeastward from Fortaleza, the low sandy hills behind the beach have scarcely a trace of vegetation that remains green through the drought periods, but between the mouth of the Assu and Cape Calcanhar lies a coast with groves of coconuts around its numerous villages. Midway in this stretch the great Lavandeira coral reef begins, a cer- tain indication that the control of muddy river water no longer takes prece- dence over the blue and saline water of the tropical ocean. Not far beyond Cape Sao Roque, in fact, and from that point southward almost to the tropic of Capricorn, the area of richest, high-sea, ultramarine surface water comes closer to the continental coast than anywhere else in America. The mouth of the Assu is close to the center of the most arid strip, with rain- fall of less than 500 millimeters a year. Over a considerable stretch of coast on either side, it is still less than 1000 millimeters, a feeble precipitation when the high average temperature and rapid evaporation are considered. Through- out northeastern Brazil the rains are all concentrated within one season, the inverno, which alternates with the dry verao. But the months of the rainy season are not everywhere the same. Recife receives 58 per cent of its rainfall during the four months between April and July; Fortaleza, 82 per cent during the five months, February to June; Joazeiro, 99 per cent during the five months, November to March. Thus we find three distinct régimes: autumn rains on the NORTHEASTERN ATLANTIC COAST 141 north coast; winter rains on the east coast; summer rains in the interior. The winds of the neighboring parts of the Atlantic explain these differences. The northeastern coast depends upon the southeast trade which here, as elsewhere, isa dry wind. It normally blows from July to December, sometimes continuing until February. Even when its rude squalls beat high against an inhospitable shore, the sky remains uniformly serene. The calms, which follow when the zone of the trade wind has withdrawn toward the south (February-May), bring the storms and rainfall. The rainfall system of Ceara recalls that of Para, with the exception that the inverno is shorter and later. The zone of calms, advancing from north to south during the southern summer, reaches Para in December or January, Maranh@o in January or February, Fortaleza in February or March. As one goes inland, the rains are earlier. The persistence of the essentially marine trade winds in the coastal region retards the formation of such barometric depressions as appear to announce the coming of summer in the interior. As for the east coast region of winter rains, it has, as on the east coast of Africa, a monsoon system. In summer (November-February) the south- east trade wind is warped to an entirely different quarter by the monsoon winds, and when it strikes the coast of Pernambuco, blowing from the northeast, it carries no rain. In winter, on the contrary, the high pressure of the continental air in northeastern Brazil causes the monsoon wind to blow from the opposite direction, that is from the south or southwest, more or less parallel with the coast. This brings heavy rains, though storms are unknown. The monsoon penetrates but a little way into the interior, which explains why the narrow, well-watered coastal zone passes abruptly into the semi-arid region (Denis, 1927). At Cape Calcanhar the northeastern corner of Brazil is turned, and the coast begins the south by east trend which extends almost to Recife. The greater part of the shore line here, and all the way down to Bahia, is a succession of white beaches and dunes, interspersed with dark green brush, clumps of coconut palms, and occasional lines of reddish bluffs. Villages are numerous, especially at the mouths of the many streams from the plateau, and there is a surprising number of relatively large towns and cities, with populations of 30,000 or more. The following three paragraphs are freely transcribed from Maull’s (1930, 9) description of the coast southward from Parahyba do Norte, a little north of Recife. From the ocean, the country seems to stretch along as a level, forested table- land, broken here and there by comparatively broad valleys. The interrelation- ships of land forms on the continental border are, however, more complicated than they appear. From the sea, the cultivated lowland is easily overlooked, not because there is little of it but because it presents no surface extensive enough to be conspicuous. Here and there one notes a line of light sandy beach. Elsewhere the dark wall of woodland is interrupted by brightly colored cliffs. In front of the forest, in some places, stands a second and lower level of vegetation, constitut- ing the mangrove zone. Finally, a reef runs along at a good distance from the 142 OCEANIC BIRDS OF SOUTH AMERICA beach, sheltering a lagoon behind a fringe of white foam. This resulting belt of quiet water has lured the inhabitarits of a coast that is poor in natural harbors not only to take to the sea but even to make long voyages in their extremely primitive craft, the “‘jangadas.’’ The latter are mere rafts, rigged with a sail and a bench. Under way they are a strange sight, with the lower structure buried in the waves, so that the boatman seems to be standing upon the water and gliding along its surface. It is said that northeastern Brazilians have gone as far as Rio de Janeiro in their jangadas, with the object of attending the Centenary Exposition. Considerable topographic variety is not lacking on this coast, for at several places the scarp of the plateau retreats inland and a broad foreland opens up. Such is the case at Pernambuco, for example, where the highlands of Olinda draw away from the shore, making room for the plain, which is a mangrove- grown, silted-up lagoon, shut off from the ocean by a tongue of land projecting from the hills. Historic Cape Sio Roque is only a slightly projecting point of white sand, about 55 meters high, with a few tufts of brush. It is visible to a distance of about 30 kilometers, and is ringed by reefs. A landing can easily be made northwest of the cape in quiet weather, and even in a fresh breeze the native jangadas run back and forth through the waves. The channel between the cape and the outer reefs is much used for local navigation. The winds at Cape Sao Roque blow generally from between southeast and east. During the south- ern monsoon season, that is from June to August, they are most southerly, and since the northwestward current then attains its greatest strength, this is the hardest time of year for doubling the cape. Between December and March the breeze is more easterly and more moderate, the sea prevailingly smooth and the current weak. Although the average breadth of both the North and South Atlantic Oceans exceeds 6000 kilometers, the distance separating South America from northwest Africa, on the shortest line from Cape Sao Roque, is only 2840 kilometers. The New World is here closer to the Old than anywhere else, except in the Arctic. This accounts in a measure for the fact that the tropical Atlantic avi- fauna is relatively uniform; the pelagic birds of the South American coast are in the main those occurring outside the shore-current zone of the African coast; in any event, there is no such division as we find between the eastern and western borders of the broad South Pacific. Such geographic proximity doubtless has a bearing upon problems even more far-reaching than those concerned with the distribution of sea fowl. For example, two species of inland tree ducks (Dendrocygna) prove to have ranges common to both Africa and South America. The same is true of the fresh-water and mangrove-swamp pochard, Nyroca erythrophthalma Wied, a duck which has been described under various specific names from widely separated localities in both of the great southern continents. The reefs that extend almost continuously around the shoulder of the con- tinent, from near the mouth of the Assu to that of the Rio Sao Francisco, are not all of coral. Perhaps more frequently they are ancient barrier strands con- NORTHEASTERN ATLANTIC COAST 143 j Fernando Noronha akoine Ascension Beck's route Murphy's route Py - *Abrolhos mallee 30\; South Trinidad [ea a Ley eee Fic. 17. Coast of eastern tropical South America and the associated oceanic islands. For detailed maps of the latter see figures 18 (As Rocas), 19 (Fernando Noronha), 20 (St. Paul Rocks), 21 (Ascension), 24 (South Trinidad and Martin Vas), 25 (St. Helena). The broken lines in the Atlantic show part of the author's course during the South Georgia Expedition. solidated by a calcareous cement, and thus forming the famous “‘stone reefs,’ which have been investigated by Branner (1904). The extent of limy precipita- tion at the edge of the ocean is explained by the debility of the coastal streams and the retention of the fresh water in the lagoons for periods long enough for it to become charged with carbonic acid and calcium carbonate. Such sandstone reefs, which have'often been mistaken for coral, accompany the Brazilian shore line, with many interruptions, from northwest of Fortaleza to Porto Seguro, beyond latitude 16° S., a distance of 2000 kilometers. The reefs are mostly straight and they usually lie across the mouths of streams and estuaries, forming natural breakwaters for the harbors behind. Wherever they occur together with coral reefs, the stone reefs are to be found on the landward side. It is possible that there may be buried coral reefs within some of them. Their geologic age is referred to the late Tertiary. The famous reef at Pernambuco is one of the stone reefs. Viewed from the sea, it looks like a long low artificial breakwater. The outer side is overgrown with seaweeds, barnacles, etc., and is also filled with chitons and sea urchins. At low tide the whole length of the reef is exposed like a black wall. (1) REEF CORALS IN SOUTH AMERICA A relationship has been indicated between the Brazilian coral reefs and the temperature, salinity, and clarity or turbidity of the coastal water. The cor- relations might easily be carried further; for the South American coast as a whole the presence or absence of reef corals is an indicator of certain qualities 144 OCEANIC BIRDS OF SOUTH AMERICA of ocean water which we have already learned to associate with definite phenom- ena in the distribution of sea birds. The location and extent of the banks and reefs of corals have been exhaustively studied by Joubin (1912), from whose monograph much of the following information is drawn. Coral reefs abound on all the islands forming the eastern border of the Carib- bean Sea, as likewise on the southern or coastal chain, from Tobago and Trini- dad westward to Curacao and Aruba. The adjacent mainland coast, however, is fringed with reefs in only a few favored localities, chiefly on northerly and northeasterly projections of the shore line, between the province of Sucre, Vene- zuela, and the Goajira Peninsula, Colombia. Farther westward, there are reefs on either side of the outer Gulf of Darien, to wit, on islets from Cartagena southwards, and along the northerly shores of eastern Panama. Since mud and corals are incompatible (as well as fresh water and marine corals), the northeastern coast of South America, between the delta of the Orinoco and Maranhd4o, is absolutely devoid of corals. Off eastern Maranhao there are a few incipient or struggling reefs and more extensive submerged coral beds. Near Fortaleza more substantial fringing and barrier reefs begin, to extend as a more or less broken chain eastward and then southward to a point just north of Cape Frio. The barrier reef along this coast is a sort of miniature of the coral growth of the east coast of Australia, which lies in but slightly more southerly latitudes. On the Brazilian coast, as in Australia, there are frequent breaks in the reef, especially off the mouths of rivers, and the inside channel is navigable in many places. The distance of the Brazilian barrier from the line of white beach is, however, mostly in the neighborhood of two kilo- meters or less. In places, moreover, the reef is so far submerged that the ocean rollers pass over it to break heavily upon the inner strand. The southernmost coral formation in South America-is a small reef on the sheltered mainland shore inside the Restinga de Marambaia, west of Rio de Janeiro and just north of the tropic of Capricorn. Off Espirito Santo and southern Bahia (16° to 21° S.), where the continental shelf widens, there are extensive coral bank formations, of which the largest is the Hotspur Bank, lying outside the Abrolhos Group. Rocas Reef and Fernando Noronha are fringed by coral growths, but St. Paul Rocks, Trindade, and the Martin Vas Islets are without them. On the west coast of South America reef corals of any sort are reduced almost to the vanishing point. In the Humboldt Current region, upwelling water produces temperatures too low to be favorable. Joubin reports insular and mainland patches in the Gulf of Panama; a short line of reefs to the northward of Buenaventura, Colombia, and some submerged banks to the southward of the same port; and reefs on the coastal ledges along the northward-facing shore of Esmeraldas, east of Cape San Francisco, Ecuador. Furthermore, he records reefs on Chatham, Charles, and Albemarle Islands, of the Galapagos Archi- pelago; at Pelado Islet in Santa Elena Bay, Ecuador; at Lobos de Tierra and Pachacamac Islands, San Gallan and the neighboring Paracas Peninsula, Peru. The reefs at San Gallan are doubtless very restricted and, as wherever else they REEF’ CORALS 145 may be present along the Peruvian coast, must owe their existence to special local conditions, such as the warm eddies so often noticeable in Pisco Bay and other indentations of the shore. 3. Tue Atiantic Equatoriau IsLaANnDs To what degree the reefs that follow the coasts of eastern Brazil afford nest- ing sites for sea birds I have been unable to discover. In the absence of any mention of beach vegetation on the barriers of either sandstone or coral, I infer that none possesses soil fully reclaimed from the ocean. We may therefore leave the continental shores for the moment in order to seek a glimpse of im- portant insular bird colonies which are, no doubt, the ultimate source of many marine species recorded from various parts of the southern tropics. a. Rocas Reef. Closest of these oceanic islands to the mainland, or at a distance of 222 kilometers, is the Rocas Reef, which is a typical coral island, and the only true atoll in the South At- lantic. Recorded dimensions are at variance, the largest figures indicat- ing that the ring is an ellipse about 8x6 kilometers in diameter. On the reef are two small cays, connected at low tide and supporting grass and a little shrubbery. Elsewhere a few knobs of weathered coral project. The highest point does not exceed 4 to 5 meters. The lagoon is shal- low, with a bottom of sand and coral, and may be entered by boats during quiet weather through a gap on the northern side. Coming northward in the brig ‘Daisy,’ I passed a short distance to eastward of Rocas Reef at sunset of April 18, 1913. From the masthead we could see one or two tumble-down buildings, the structural support of the automatic light, and a single tall coconut palm, the sole survivor of three which are said to have been planted in 1857. No sign of the island bird life was visible save for a few boobies in the air far off. Rocas was later visited by the ‘Blossom’ party, during the South Atlantic Expedition of the Cleveland Museum of Natural History. This was also in April, and Simmons (1927, 71) found hosts of birds breeding. Sooty Terns (Sterna fuscata) he estimated to number “‘hundreds of thousands.’’ They were sitting so close that no indication of the horde was apparent from a distance, but a cloud of birds arose from the eggs when the colony was invaded. On one of the cays Simmons estimated that there were also approximately 1200 nests ROCAS REEF Miles i 2 Kilometers. ! Heights in meters gn f Sa Fic. 18. 146 OCEANIC BIRDS OF SOUTH AMERICA of the Brown Noddy (Anods stolidus), 1500 of the Blue-faced Booby (Sula dacty- latra), and a few of other birds. Furthermore, he observed about the island some 350 individual examples of the Brown Booby (Swla leucogaster), 25 of the Red-footed Booby (Sula sula), and 15 of the Man-o'-war Bird (Fregata magni- ficens). The last three species may likewise have been breeding, though sites would necessarily be scarce for the Red-footed Booby, which in all parts of the world apparently requires trees or bushes as a support for its relatively well-constructed nest of twigs. Between the schooner and the reef Simmons saw both Wilson's and Leach’s Petrels, the former a northbound migrant from the far south, the other a northern-winter visitant close to the southern limit of its pelagic range. All in all, the resident avifauna of Rocas Reef is little if any different in character from that of certain offshore islets of the Caribbean, or of outliers of the Antilles. b. Fernando Noronha. About 134 kilometers east of Rocas Reef, or 356 from the Brazilian coast, lies the burnt-out volcanic island of Fernando Noronha. It is in reality a group, the lesser islets stretching toward the northeast being called, respectively, Platform, Egg, St. Michael's Mount, Booby, and Rat. Off the southern side is an isolated pinnacle. Although some of the smaller members are partly of sandstone conglomerate, the geological features of Fernando Noronha as a whole are closely akin to those of South Trinidad, which lies more than 1600 kilometers to southward. Clumps of coral grow on some of the rocky shores and on neighboring shallow parts of the sea bottom, though there are no actual reefs. The main island, which measures about 7 by 3 kilometers, has an undulating plateau rising to 100 meters or so above sea level. Near the middle of its north- erly coast there is a bold, overhanging, phonolitic dyke known as the Pyramid, which is upwards of 330 meters in altitude and is the most arresting landmark in the whole South Atlantic Ocean. On the level terrain the soil is thick and rich as a result of the accumulation of humus during a long period. Fernando Noronha lies in an ocean as warm as the Caribbean. During the season of northern-hemisphere summer it is swept steadily by the southeast trade winds, which are more or less interrupted by doldrum conditions during the northern-winter season. Between February and mid-July the island is well watered; numerous streams form, and waterfalls tumble from the luxuriantly covered cliffs. The mean annual precipitation amounts to 1294 millimeters, but the total is extraordinarily irregular and has varied within a twenty-year period between 460 and 2396 millimeters. Rain is normally scanty during the last four or five months of the year, but the climate is rarely oppressive. It will be noted that here, as on the mainland, the windy season is the dry season. When the trades are strongest, a spectacular surf pounds the windward coasts of Fernando Noronha. During the season of calms, on the other hand, the more sheltered shores are singularly peaceful. Curving beaches of sand run down to ATLANTIC EQUATORIAL ISLANDS 147 a | FERNANDO NORONHA Miles ‘ Kilometers 6 i z 3 Heights in meters CH rcty 1. Patho CS St Michael’s Mount JS Antonio Bay a pice = The Twins Qo cS Ollha dos Sinos tee laay OLesPattes g S OSI. das Fragatas “nes, ih I. - Tobacco or South Pt Fic. 19. quiet, pale green coves, the hue of which is reflected from the white breasts of sea birds flying overhead. The majority of the oceanic fowl appear to have no fixed breeding season, for the eggs and young of a number of species can be found during any month of the twelve. The Sooty Tern, however, forms an exception to this; here as elsewhere all members of this species apparently lay their eggs together, even though the date of the nesting period may vary somewhat from year to year. The bulk of the sea bird population, particularly the ground-nesting species such as the Sooty Tern, the Brown Noddy, two of the resident boobies, etc., is not on the main island but rather on the islets. The outermost member of the chain, Rat Islet, is upwards of 3 kilometers in length and is rugged and precipi- tous. It has been worked sporadically for calcium phosphate derived from ancient bird guano. Estimates recorded in the pilot books allege that half a million metric tons of this substance remain to be extracted. Fernando Noronha was originally heavily forested. At the time of Darwin's visit, in February, 1832, the main island was largely solid woodland, so thickly intertwined with creepers that to crawl through it required great exertion. Darwin was also impressed by the large ‘‘magnolias,’’ by trees covered with delicate pink flowers (doubtless Jatropha gossypifolia), and by the general beauty of the scenery. He expressed disappointment, however, that the grandeur of the island by moonlight rather faded at dawn, when the hills seemed to shrink under the light of day (Darwin, 1933, 37). He also likened the island (J. c., 341) 148 OCEANIC BIRDS OF SOUTH AMERICA to the drier parts of the Galapagos, a comparison which seems all the more apt since we now know that the latter archipelago, no less than Fernando Noronha, is characterized by great irregularity in the amount and seasons of rainfall. The character of the vegetation has been greatly changed through long Occupation of the island as a penal colony, and through the introduction of familiar tropical culture plants, such as the banana, coconut palm, maize, etc. Nevertheless, Fernando Noronha still supports a dense and varied flora of indige- nous trees, shrubs, vines, and cacti, a number of which are endemic forms, as are likewise several of the insects, lizards, and birds. Most noteworthy among the reptiles is the native legless lizard (Amphishaena ridleyi). This animal has been made the prop of far-fetched land-bridge hypotheses, which would make of Fernando Noronha the relict of a continuous terrestrial path between Africa and America. It is, however, closely related to a West Indian form, making a link with the Antilles more notable than any evidence exhibited by the birds. Two of the three resident land birds are endemic. The trio comprises a race of a widely distributed South American dove (Zenaida auriculata noronha) which is found in northeastern Brazil as well as on the island, a tyrant flycatcher (Elainea ridleyana), and a viteo (Vireosylva gracilirostris). The dove forms some- what odd associations with the sea fowl, particularly when the habits of its continental relatives are taken into consideration. Moseley, naturalist of the ‘Challenger,’ found it nesting among the boobies and noddies on ledges along the low cliffs of Booby Islet, the eggs and young of the three species being closely intermingled (Moseley, 1879, 83). Fernando Noronha has been frequently, although scarcely more than casu- ally, investigated by numerous naturalists since early in the nineteenth century. A bibliography of the more important scientific accounts accompanies a report on my own brief visit in 1912 (Murphy, 1915, 46). The following summary of the sea bird fauna is derived from the earlier literature, from my personal observations during a single day's landing in October, and from the subsequent visit of the schooner ‘Blossom,’ made in March and early April (Simmons, 1927, 60). The breeding species include two tropic-birds, three boobies, one man-o’-war bird, and four terns. They are as follows: Phaéthon aethereus ~ Fregata magnificens Phaéthon lepturus Sterna fuscata Sula dactylatra Anois stolidus Sula sula Anotis minutus Sula leucogaster Gygis alba Fernando Noronha appears to mark the southernmost breeding station on an oceanic island of Fregata magnificens, which is the characteristic man-o'-war bird of the northern tropical and sub-tropical Atlantic-Middle American zone. Its specific range extends all the way from the Cape Verde Islands on the east to the Galapagos on the west. In north-south extent it stretches from the Bahamas to eastern or southern Brazil, for this same species nests on islets close ALEANTIC EQUATORIAL ISEANDS 149 along the coast of the mainland to southward of Cape Sao Roque. All the specimens of man-o’-war birds I have seen from Rio de Janeiro and other parts of the southeastern Brazilian coast are magnificens. Ribeiro (1919, 186) and Peters (1931, 97) have listed such littoral birds as belonging to the species minor, which inhabits South Trinidad, but this appears to be incorrect. Ascen- sion Island, in the middle of the Atlantic and only four degrees farther south- ward than Fernando Noronha, is occupied by the endemic and highly sedentary, species, Fregata aquila, while the larger of the two forms of man-o’-war birds inhabiting South Trinidad, more than fifteen degrees of latitude to southward is a representative of Fregata minor, which has a world-wide distribution in the southern tropics and sub-tropics. Of the other Pelecaniformes resident at Fernando Noronha, the Red-billed Tropic-bird (Phaéthon aethereus) has a range most nearly resembling that of Fregata magnificens. This tropic-bird is likewise found throughout the Atlantic- American region, from the Cape Verdes to the Galapagos. But it has also penetrated southward to Ascension and St. Helena, in addition to which a race of the species occurs in the Persian Gulf and the northern Indian Ocean. The Lesser Tropic-bird (lepturus), the three boobies, and the four terns all belong to species having a more or less world-wide tropical distribution. Transferring our attention to birds belonging chiefly to the southern tropics, it is highly interesting that Fernando Noronha marks the most northerly and easterly breeding station in the Atlantic of the Fairy Tern (Gygis). The resident form is an isolated South Atlantic race, inhabiting only the four islands of Fernando Noronha, Ascension, St. Helena, and South Trinidad. Elsewhere in the milder zones of the southern hemisphere, namely throughout the Indian and Pacific Oceans, the white Fairy Terns have as nearly a continuous distribu- tion as is possible for insular birds which normally do not stray far from their breeding sites, and which consequently, like the man-o’-war birds, tend to become split up into numerous geographic races. Since there are seasons when powerful southeast trade winds blow from Fernando Noronha across the equa- tor almost as far as the mouth of the River Orinoco, speculation offers me no clue as to why Gygis has not succeeded in jumping the next gap and establishing itself in the West Indies. Considerably greater distances in the Pacific have been crossed by birds of this genus. Fairy Terns transported to certain islands of the Lesser Antilles would be as assured of a congenial environment as would the previously discussed Brown Pelicans if introduced into southern Brazil. Finally, Fernando Noronha is noteworthy in that it appears to have no native petrel of any sort. One might assume that its rocks and soil, and the waters of the surrounding ocean, would be well suited to the resident needs of Audubon's Shearwater (Puffinus lherminieri), which inhabits the Antillean-Gulf of Mexico region, the Bahamas, and even Bermuda; a distinct race, moreover, is found in the eastern Atlantic at the Cape Verdes. The West Indies have also one or two other native petrels, while no less than six species breed among the islands of the Cape Verde group. Ascension, St. Helena, and South Trinidad each support at least one species. 150 OCEANIC BIRDS OF SOUTH AMERICA c. St. Paul Rocks. Far to the northward and eastward of Fernando Noronha, and at more than double the distance (869 kilometers) from the South American mainland, lies the little horseshoe-shaped cluster known as St. Paul Rocks. The longest axis, from the southernmost to the northernmost rock, is scarcely more than 300 meters, and the highest peak, which is white with the droppings of sea birds, is not more than 20 meters above the level of the sea. Nearer the water the rocks are greenish black. Most of the outliers are more or less mushroom- shaped because the ocean has sapped and undermined them. There ts no ter- restrial vegetation larger than unicellular aigae. St. Paul Rocks are rare if not unique among oceanic islands in that they are neither volcanic nor coral formations. They are composed entirely of ultra- basic igneous rock of plutonic aspect, and ST. PAUL ROCKS represent perhaps a summit squeezed up by ant lateral pressure from the floor of the Atlantic ata Shae apenas Ocean. Darwin (1933, 35) wrote during the Meters cruise of the ‘Beagle’ that a mile from the pe eccrgutin islets the lead could not find bottom. Sound- Se ares ings made by the ‘Quest’ indicate that ‘at no 5 Pillar Rock 6 point is the 100-fathom line distant more than four cables from the rocks, and in places is within 100 feet’’ (Wild, 1923, 80). More re- cent sonic measurements by the ‘Meteor’ show that the rocks represent the pinnacles of a steep-sided mountain no less than 4000 meters in height (Spiess, 1928, 78). They lie in an ocean which is prevailingly stormless and cloudless, except during the brief tempests of # Lat, 0°55'28’N, Long. 25°22'32’W.| thedoldrum season. During even calm weather, however, they are usually beset by strong surf, which splashes them with spray to their sum- mits and tends not only to wash away the guano of the sea birds but also to cause it to precipitate its salts in the form of a dense, shining, glassy layer, which has been described by Darwin, Sir James Clark Ross, the ‘Challenger’ naturalists, and other explorers. According to Moseley (1879, 72) this enamel- like crust is hard enough to scratch glass. It is here given more than passing mention because a similar encrustation is found upon one or more islands off the west coast of South America, to northward of the arid zone. Despite the difficulties of landing, St. Paul Rocks have been visited by a con- siderable number of oceanographic expeditions. It was here that the youthful Darwin, on February 15, 1832, first found himself in the midst of a colony of sea birds and, fired with the enthusiasm of a collector, began to knock down examples with his geological hammer. So far as I can determine, the resident birds are confined to three species, namely, the Brown Booby (Sula leucogaster), and both of the Atlantic noddies (Anois stolidus and Anois minutus). Boobies Fie; 20: ATLANTIC EQUATORIAL ISLANDS 151 thickly cover the higher slopes and summits. Their food consists largely of flying fish, with which visitors have found these birds so crammed that they seemed constrained to disgorge before taking flight. Flying fish brought ashore by the boobies, or inadvertently blown on the rocks by the wind, also supply much of the food of the amphibious crabs which teem on all of the islets. However, the crabs also devour young birds, and Moseley was amazed at the apparent ingenuity of these crustaceans in circumventing their supposedly more intelligent vertebrate rivals. Of the noddies, the larger and brown species (Anois stolidus) lays its single egg on bare rock, chiefly in situations lower down than the booby sites. The lesser or Black Noddy has quite different breeding habits. This species is in most places a tree-nesting tern, but at St. Paul Rocks it builds compact, bracketed nests of green seaweed, cemented with dung, on small projections of the slopes and ledges. It is a less abundant member of the island bird life than the other noddy. A few insects, mites, and spiders complete the terrestrial fauna of St. Paul Rocks. The cove embraced by the islets abounds with sharks and with a great wealth of tropical rock fishes. Among surface species the flying fish have al- ready been mentioned. What kinds of small fry support the noddy population has not yet been reported; perhaps there are enough larval flying fishes to answer the purpose. Noddies do not plunge for their prey in the manner of certain other terns, but rather seize it in mid-air as the tiny, minnow-like fishes leap or sail from the water. Some of the books on nautical information still give directions, based upon the experience of that redoubtable voyager, Captain Amasa Delano, in the year 1799, for procuring birds’ eggs for food at St. Paul Rocks during the month of November. More up-to-date ornithological information demonstrates, how- ever, that these equatorial islets have a characteristic oceanic régime, without a restricted breeding period for the birds, and possibly without even a seasonal peak of productiveness. Moseley, of the ‘Challenger,’ wrote (1879, 73), ‘Fitz Roy visited St. Paul's Rocks on February 16th; Ross on May 29th; we on August 29th; on all these occasions eggs and young birds were found. Hence, breeding goes on all the year round."’ The experience of other parties, such as those of the ‘Meteor,’ in early May, of the ‘Quest’ in October, and of the ‘Valhalla,’ in December (Nicoll, 1908, 2) amply confirms Moseley’s deduction. d. Ascension. The last of the islands in the Atlantic tropical zone that needs consideration is Ascension, which lies approximately halfway between the shoulder of Brazil and the Lower Guinea coast of Africa. Ascension is a youthful volcanic island, or a ‘‘cinder of the sea,”’ nearly circular, with a maximum diameter of 12 kilo- meters and an area of 98 square kilometers. From the summit of Green Moun- tain, 859 meters above the sea, about forty lesser hills, representing former vents, may be counted. Most of the natural vegetation of the island is confined to the cap of the mountain, which receives considerably more moisture than 152 OCEANIC BIRDS OF SOUTH AMERICA ASCENSION ISLAND Miles 2 Kilometers —_— North Pt ' 2 3 “79 /, 4 fg Heights in meters Oy Porpoise Pt x North|East Pt Hummock Pt OBoatswain-bird |. Dampiers Spring ° 599 s G mtn South East Head eee Heaton ‘air MSArthur Pt aS South West Bay 378 Portland Pt Fie. 21. the lowlands. The annual average precipitation amounts to only 500 milli- meters, the greater part of which falls during the months of March and April. Many species of culture plants have been introduced into Ascension, and are now thriving there, particularly in gardens on the higher slopes of Green Mountain. The introduced animals include a number of birds, such as Guinea fowl and African weavers. For about a century after the date of Napoleon's exile at St. Helena, Ascen- sion was operated as a British ship of the line and was known and governed as ‘H. M.S. Ascension,’ but its administration has within recent years been turned over to a commercial cable company. The ornithological interest of Ascension is largely associated with the fact that the island is the type locality of no less than four species of ocean birds described by three early naturalists. These birds are: Phaéthon aethereus Linnaeus, 1758 Fregata aquila (Linnaeus), 1758 Sula dactylatra Lesson, 1831 Gygis alba (Sparrman), 1786 All of these have specific or subspecific representatives in warm seas through- out the globe, and the precise determination of the forms residing at Ascension ATLANTIC EQUATORIAL ISLANDS 153 is the cornerstone of much taxonomic study. As regards the Man-o'-war bird, for example, no other species of this genus was generally recognized by ornithol- ogists for nearly a century after Linnaeus had named Fregata aquila. Conse- quently much of the literature relating to four or more species of Fregata is listed under the specific name aquila. In point of fact, Fregata aquila is now known to be a bird of such restricted distribution that it has never been taken anywhere except at Ascension. Until recently, Ascension was also held to be the type locality of a fifth sea bird, the Red-footed Booby (Sw/a sula Linnaeus 1766). Grant and Mackworth-Praed (1933, 185) have recently shown, how- ever, that this should properly be assigned to Barbados, of the Lesser Antilles. Ascension has apparently been the home of great colonies of sea birds through- out past ages, for it is reported that a fertilizer company has recently been con- ducting operations at the northern corner of the island, where the cinder plains come down to the sea. In pits of the lava and volcanic ash there are here quan- tities of phosphate. As Simmons (1927, 73) remarks, it would be of interest to search for bird bones in these deposits, for the producers of the ancient guano are not certainly known, and they may have been different in part from the modern birds. Even within historic times sea fowl were undoubtedly more numerous on the main island than at present. Thus Darwin (1933, 415) wrote in his diary for July 22, 1836, that he could not imagine the cause of the white patches with which the whole plain was mottled, as seen from the deck of the ‘Beagle.’ Subsequently he found that ‘‘it was owing to the number of seafowl, which sleep in such full confidence, as even in midday to allow a man to walk up & seize hold of them."’ The description suggests white boobies, although a view of Sooty Terns directly to leeward, with all the birds consequently facing the wind and the observer, might also appear prevailingly white. The principal present-day headquarters of sea birds are in two places, namely, (1) the Wideawake Fair, which is the famous colony of Sooty Terns at the western corner of the clinker plain; and (2) Boatswain-bird Islet, a short dis- tance off the eastern coast of the main island. Both localities have recently been visited by the ‘Blossom’ party of the Cleveland Museum of Natural History (Simmons, 1927, 44). Boatswain-bird Islet rises steeply, glaringly white with guano, against the dark volcanic mass of Ascension. Its ledges and talus slopes are occupied by Sula dactylatra, and its plateau by both this species and the Man-o’-war bird. In the absence of vegetation, the latter here nests directly upon the ground, as the species Fregata magnificens likewise does at St. Michael's Mount, Fernando Noronha. Simmons lists also the Brown and the Black Noddies, the Greater and Lesser Tropic-birds, the Fairy Tern, and a Mother Carey's chicken (Oceanodroma castro) as residents of Boatswain-bird Islet, despite the fact that it possesses neither a sprig of green nor a drop of water. Wideawake Fair has been described by practically every visitor to Ascension. During the breeding season, the eggs of the Sooty Tern or Wideawake, are so thickly distributed that it is difficult to traverse the district without stepping upon them. As many as 10,000 dozen edible eggs have been taken from the colony within a week, but since the robbed pairs of terns immediately lay 154 OCEANIC BIRDS OF SOUTH AMERICA again, there seems to be no diminution in their numbers from this exploitation. The observations of the ‘Blossom’ party seem to confirm the time-honored story that the nesting season of the Sooty Terns begins a little earlier in each succes- sive year. The change in time is sufficient to make the birds breed, on the average, four times within three years. To the sea birds thus far listed as breeding at Ascension, we should add the Brown Booby (Sula leucogaster), which nests on Boatswain-bird Rock, where specimens have been collected by members of several expeditions. This makes a total of eleven certain residents, provided the Red-footed Booby (Sula sula) can be so regarded. The latter species seems to be invariably a tree-nesting or bush-nesting bird, and arborescent vegetation is altogether lacking at Boat- swain-bird Islet and extremely scarce on the main island. Osbeck (1771, 2, 78), who explored Ascension during April, 1752, when it was still uninhabited, states that it was even then ‘‘without woods.’’ However, he adds that the island had “‘formerly had woods, as appears from several perfect petrefactions of branches of trees, and pieces of wood; but in particular from a large petrified stump'’ (81). As a further indication of a native flora, Kinnear (1934, 32) has recorded notes and a drawing from the manuscript journal of Peter Mundy to show that an otherwise unknown species of flightless rail inhabited Ascension within historic time, 2. e. cérca 1655. No specimens of the Red-footed Booby are listed among the birds obtained during many historic British expeditions (Ogilvie-Grant, 1898, 434), nor did the Cleveland Museum party discover the species at Ascension. Nevertheless, Mathews (1915, 216) once designated Ascension as a breeding station of the Pelecanus Sula of Linnaeus, upon the apparently sound testimony of Osbeck @. c., 89), which is as follows: Diomedea Adscensionis was caught here. It was entirely white, not even the thirteen feathers in the tail excepted; had red feet, formed chiefly for swimming; and only black tips to its wings: for the rest, it is like the Diomedea ptscatoria This description could probably apply to no other bird, and we may con- clude that the Red-footed Booby was a former resident, whether or not it still breeds at the island. Furthermore, Osbeck first noted the presence of the nesting petrel (Oceanodroma castro), examples of which have only recently been collected at Ascension. He writes (89): ‘‘We also saw a species of little black sea birds, but only upon the wing.” Finally, there is the dubious case of the southern tern, Sterna vittata, which is listed as a breeding bird of Ascension by a number of recent authorities (Alexander, 1928, 169; Peters, 1934, 333). Careful search has thus far failed to yield the source of this record. In other parts of its range the tern is confined to sub-antarctic or to barely sub-tropical islands, and I suspect that the inclusion of the species as a breeding bird of Ascension, or even of St. Helena, is merely an oft-repeated error. Sclater (1924, 149) is more cautious than some of the other compilers; he states only that Sterna vittata ranges in the South Atlantic and Indian Oceans from Ascension to Tristan da Cunha and Kerguelen Island, ATLANTIC EQUATORIAL ISLANDS 155 nesting at Tristan. The whole substance of the original association of this tern with Ascension may be sought, perhaps, in a misinterpretation of Reiche- now's (1908, 562) statement: ‘‘Streicht nordwarts bis zur Gough-Insel, Tristan d’Acunha und ist sogar auf der See zwischen St. Helena und Ascension erlegt.”’ (1) ROLLERS Ascension, like several of the other Atlantic islands, is notable for the phenomenon of ‘‘rollers,’’ or great waves, which are not related to movements of the surface caused by the prevailing southeast trade wind. They are said invariably to break with particular force on the leeward side of the island. Such waves occur perhaps equally at St. Helena, South Trinidad, and Fernando Noronha. The late Lady Gill, who accompanied her husband to Ascension in July, 1877, before a historic measurement of the solar parallax, wrote as follows of the rollers during October of the same year (1878, 147): During the five previous days they had been persistent, and for the first twelve hours their grandeur and power exceeded anything I had ever conceived. I thought I had seen rollers at their worst on the day we landed at Ascension, and again on the night of the eclipse, but these I now found were but baby rollers after all. The full-grown giants shook our little encampment like an earthquake, and the noise of their thunder deafened us. What a sight it was! My pen is quite powerless to describe it. They fascinate one, too, these mysterious rollers, and, watching them, we enjoyed our evening stroll along the shore more than usual. Yet, each time that a great wave rosé up twenty or thirty feet high, and came thundering along to dash itself to pieces on the beach, I shrank back with a sort of involuntary desire to flee the sight. According to Meliss (1875, 392, 404) the rollers usually reach Ascension from one to seven days before the same rhythms strike St. Helena. At both islands their direction of progress is southerly or southeasterly, so that they break only against northerly shores. At St. Helena the worst manifestations occur after prolonged lulls in the trade wind, between December and March. The long-remembered rollers of February 17, 1846, appeared during early morn- ing, without other warning than the calm and sultry air, and, by evening of the same day, thirteen ships at anchor in the harbor of Jamestown had been shattered to bits against the rocks. Waves of this nature take their origin from distant meteorological disturb- ances, and have nothing to do with winds prevailing locally. In some instances they may represent hurricane waves, such as have been described on p. 55. Their source, on the other hand, may more often be due to much less perceptible stimuli. It is now known, for example, that similar great swells breaking in the neighborhood of Casablanca, on the coast of Morocco, are caused by baro- metric depressions over the ocean between Iceland and the Azores. The rollers of the South Atlantic islands, and the swells known in the neighborhood of Rio de Janeiro as ‘‘resacas,’’ are doubtless equivalent phenomena, as are also the “‘surf days’’ which frequently come during calm weather at the Chincha Islands and other historic guano-loading localities along the Pacific coast of the continent. In former times the shipmasters who contracted to transport 156 OCEANIC BIRDS OF SOUTH AMERICA guano always demanded a special allowance for a certain number of days during which the mysterious and unpredictable rollers from the open Pacific would entirely prevent them from carrying out their accustomed labor. The birds of the Peruvian guano islands, and no doubt the cliff-dwelling sea fowl of the South Atlantic islands, also take unconscious account of the conditions, for rarely if ever are nesting sites selected on ledges exposed to the capricious fury of the rollers. 4. RectFE DE PERNAMBUCO TO THE Rfo DE LA PLATA To southward of Cape Sao Roque there are more navigable rivers than to northward and westward, and the entrances to many of the larger inlets possess effective natural breakwaters in the stone reefs. All this part of the coast, however, to a point south of Bahia, lies within the region of anomalous south- ern-winter rainfall, and the long dry season causes such a shrinkage in the amount of running water that some of the streams can be entered by vessels only during the rainier parts of the year. The point emphasizes merely the unequal distribution of rainfall for, upon the basis of figures, all of this coast should be regarded as well watered. The mean annual precipitation at Parahyba amounts to 1763 millimeters, at Recife 2092 millimeters, at Bahia 1876 milli- meters. Nevertheless, away from the river deltas, the flat shoreline strip is generally sandy, and covered, even where moisture is copious, with a some- what scanty vegetation in which coco palms predominate. Cape Branco, close to latitude 7° S., from where the coast runs nearly due south, may be taken as the easternmost projection of the continent. To south- ward of the cape, the reef, which is usually studded here and there with the hulks of wrecks, is particularly fragmentary, or even missing entirely along considerable stretches of shore. For some distance, moreover, the mainland coast is here made up of bold and conspicuous reddish cliffs, visible from far at sea. Such islands as are found along this part of Brazil are obviously delta or estuarine formations. The largest is Itamaraca, a little north of Recife, and lying in a bight of the mainland from which it is separated by a narrow channel. Its seaward shore is covered with groves of coconut palms, among which are the white dwellings of fishermen, and the fronting beach is constantly pounded by breakers from the open Atlantic. At Recife itself the remarkably artificial looking stone reef, which has given the port its name, thoroughly protects the roadstead from the battering of this same surf. From the neighborhood of Recife the coast line begins a. southwesterly trend that extends to the Bahia de Todos os Santos; likewise it grows more irregular and picturesque. Beyond the mouth of the Rio Sado Francisco, hills become visible from the ocean; the land is higher and there are many small bays and headlands to replace the monotonous sandy beaches that extend from Maranhao to Recife (Smith, 1879, 1, 448). The significance of this change is that the sedimentary zone has narrowed until the granite massif touches the coast. Between Parahyba and Bahia the table-land rises from 60 to about 200 meters, CENTRAL ATLANTIC COAST 157 while the valleys, on the other hand, take on toward the south a more sub- merged appearance, until their outer parts resemble fiords. Some of them are filled up with mangrove-covered alluvium, while others are blocked off by a continuous barrier beach, behind which is a lagoon or an ill-drained swampy tract with its long axis at right angles to the shore line. Near Maceid, for example, are two such lakes, fed by rivers from the interior of the country and communicating through narrow mouths with the ocean. Such bodies of water are prophetic of the still more extensive lagoons to be found farther southward on the Brazilian coast. The Bay of Todos os Santos at Bahia is itself a drowned estuary which, however, maintains a wide connection with the sea. Wherever along this coast the Tertiary beds of the table-land are worn away by the waves, they form straight cliffs which furnish abundant material for the coastal cur- rents to transport. In favorable places, therefore, flats have secondarily formed in front of such precipices, making young foreshores now grown over with coconut palms. To northward of Bahia, the original coastal forests have mostly been de- stroyed, owing to the relative density of human population mentioned above. For a long distance south of Bahia, however, forests practically everywhere cover the eastern edge of the continent. The rounded, tree-clad mountain ranges are visible from the ocean long before the low plains at their bases appear. To northward of Rio de Janeiro such plains are in the main due to breaks in the ranges, which allow the rivers to carry their detritus into the Atlantic. South of Rio, where the maritime mountains are more nearly continuous, narrow coastal plains of this type are mostly lacking. At Ilheos, a little to southward of Bahia, the southern limit of the region of anomalous winter rainfall is passed, and we reach a part of the coast in which the dry season is little marked. The annual precipitation amounts to 2234 millimeters. The bulk of the rain comes between November and May, but there is a sceondary maximum in September. From this point southward the predominance of summer rainfall becomes more definite. The six months of southern winter account for only 22 per cent of the total annual rainfall at Campos, inland from Cape Sio Thome. At Rio de Janeiro the annual rainfall of 1118 millimeters is more equally divided, 36 per cent being received during the six winter months. From Bahia the coast runs nearly southward for 400 kilometers, or as far as the vicinity of the Abrolhos bank and islands. It is a fertile and well-settled stretch, with numerous modern Brazilian towns as well as villages still in- habited by the aborigines. Broad sandy beaches alternate with wooded hill- ocks, the latter sometimes carved into reddish cliffs toward the sea. The mountain barrier is not high along this part of the shore, the peaks ranging up only to about 600 meters. Mouths of rivers can be distinguished by the breakers on the reefs that protect most of them but, owing to the general absence of striking landmarks, many impermanent objects, even such as conspicuous trees, have come to receive more than ordinary recognition from mariners. Along parts of this coast coral growths run close to the shore; elsewhere 158 OCEANIC BIRDS OF SOUTH AMERICA deep water extends to the very base of the continental scarp. The northern edge of an extensive submarine plateau, upon which lie the Abrolhos Islands and reefs, begins at about latitude 17° S. The bank measures roughly 250 by 200 kilometers, is as shallow as 30 or 40 meters over wide areas, and yet rises abruptly on its offshore side from abyssal depths. Between the Abrolhos and the outlying Rodgers Bank, for example, the ‘Meteor’ obtained echo-measure- ments of 3800 meters, and along the southeasterly side of the main bank found a very steep slope descending to 2000 meters (Spiess, 1928, 233). Along the margins, soundings of 60 to 600 meters are in many places in close proximity, and no less than ten or twelve small banks, separated from the principal plateau by deep water, approach close to the surface. Thus the Abrolhos Bank and the outlying fragments really form a considerable chain of mountains under the sea, with five peaks of no great extent projecting into the atmosphere. The islets are constructed of the familiar Brazilian reef lime, rather than of coral, but they are fringed and surrounded by coral formations. There are four principal members and two smaller ones, arranged in an irregular ring, and lying approximately 65 kilometers from the nearest point of the mainland. The five named islets are called Santa Barbara, Redonda, Siriba, Southeast, and Guarita. Santa Barbara, the site of the lighthouse, is in latitude 17° 57’ 31’’ S., and is 33 meters in maximum altitude, with a length of slightly over a kilometer. The vegetation comprises only cactus, and small shrubs. All of the Abrolhos Islets are said to be covered with the nests of innumerable sea birds about which, however, little appears to be known. The only specimen that I have been able to find among many collections of oceanic birds is an example of Sula dactylatra. Doubtless other species of boobies, frigate-birds, Sooty Terns, and one or both of the Atlantic noddies make up a good proportion of the resident sea bird population. Darwin (1933, 46) gives a record of a landing from the ‘Beagle’ at the Abrol- hos Islets on March 29, 1832. This was marked by a larder-raid upon the sea fowl, of which such an enormous number were slaughtered by the aid of guns, sticks, and stones, that the victims were more than the ship's boats could hold. Unfortunately, no examples appear to have been preserved, nor are the kinds named. On the sea at a distance from the islets, or around the borders of the bank, Darwin reports that the only birds seen were Mother Carey’s chickens. The latter were probably Oceanites oceanicus, numerous specimens of which were collected by Beck close to the Brazilian coast at the same'season of the year but nearly a century later. The Abrolhos are the seat of an extensive fishing industry carried on by the inhabitants of the neighboring coastal villages. Everywhere about the islets the water is shallow, warm, and clear, and the coral reefs of both the fringing and the barrier types harbor multitudes of rock-dwelling and bottom fishes, besides which flying fish and other surface species come close inshore from the Brazil Current. While the upper waters of the open ocean in this region are in general too warm to support a teeming marine life, the very presence of such islets as the Abrolhos has an enriching local effect. This may be due in part CENTRAL ATLANTIC COAST 159 Fic. 22. Bathymetric chart showing the continental shelf (to 200 meters), the oceanic ridges and plateaus bounded by the 4000-meter contour (light areas), and the deeper basins of the South Atlantic and eastern South Pacific Oceans. Note that the Whaleback Ridge connects the African and mid-Atlantic plateaus. From the American Museum homolographic base map of the world (1931). to the forcing upward of deeper and cooler waters by the slope of the bank. According to Schott (1926, 109), both the ocean waters and the surrounding land of what he terms the ‘Brazilian Basin’’ are prevailingly cooler than the water and atmosphere of the same latitudes on the African side of the Atlantic. In the western bed of this ocean there is nothing corresponding to the Whale- back Ridge, which bars the cold and deep polar waters out of the African Bight. The northern limit of the Brazilian Basin extends from St. Paul Rocks west- southwestward, by way of Fernando Noronha and the Rocas Reef, to Cape Sao Roque. Thus only an interrupted line of submerged mountains corresponds with the Whaleback off the African coast. In the neighborhood of the Abrolhos Islets there are interesting coral struc- tures differing from the several types of ordinary reefs and banks. Most notable among these are the clumps or coral-heads, which the Brazilians call ‘“‘chapei- rdes."’ These are small patches in the open sea, which rise like pillars from depths of 15 or 20 meters or more, their tops sometimes attaining the level of low water. Occasionally they spread out aloft, like mushrooms. The Parcel 160 OCEANIC BIRDS OF SOUTH AMERICA dos Abrolhos, on which many vessels have been wrecked, is a cluster of such coral structures to eastward of the archipelago and several square kilometers in area. In quiet water, and especially when viewed from a distance, the surface of the sea over the chapeirGes appears to be flecked by shadows, each of which marks the position of a coral column. To southward of the Abrolhos Islets the coast of Espirito Santo is generally low, with spits and sandhills dividing the sea from the lagoons and marshes that extend inland 40 kilometers or more to the base of ranges parallel with the shore. The Rio Doce, notable among Brazilian rivers for the fruitfulness of the land it traverses, enters the Atlantic at 19° 37’ S., with very extensive banks, formed by river detritus, off its mouth. The flotsam includes trunks of large trees, many of which stick at various angles in the changing bars, while during the rainy season others may be encountered floating at sea, far off the mouth of the stream. To southward of the Doce the ocean encroaches still more closely upon the elevated crustal block of the continent, which is bounded by scarps. The fore- land is higher, with increasing slopes and cliffs, and peaks with an altitude of 1000 meters or so are visible from the ocean. Rounded islets, with brushy or wooded tops, become more and more common close to shore toward Cape Frio and Rio de Janeiro. Evidence of a double geological shift here begins to be indicated. The first movement involved a recent slight uplift, shared with more northerly sections of the Brazilian coast, as well as with Patagonia; the second a still more recent and more pronounced submergence, increasingly marked toward the south. The sunken bays of Rio, Santos, and Paranagua all testify to this depression of the immediate shore. The great bend in the coast just to eastward and northward of Rio is formed by two salients, of which Cape SAo Thome, the more northern, is low and sandy, with a vast swampy plain behind it, while Cape Frio is bold and double- summited, with wooded slopes and an altitude of about 500 meters. Cape Frio is often the first landfall for ships bound toward Rio from the north and east, and here clouds of butterflies, or bewildered land birds perching in the rigging, are likely to extend a welcome to tropical lands. From Cape Frio, the shore, rising steadily westward from strands to bluffs and cliffs, leads to the harbor of Rio de Janeiro. The best general view of the beautiful site of the city is to be obtained from the ocean to southward. Thus seen, the peculiar and much- described topographic formations within the metropolitan area form a united picture with the Organ Mountains rising behind. Within the harbor is an island called Cagarra, a word suggestive because in Old-World Portuguese settlements, such as the Azores, Madeira, and the Cape Verde Archipelago, it is the name given to a large shearwater (Puffinus). However, the term connotes merely the conspicuous defecations of sea birds on the rocks, and the islet at Rio is more likely to be an ancient breeding place of the Brown Booby than of any species of petrel. Along the southward-facing coast on both sides of Rio de Janeiro there is an eddy-like countercurrent, called the Sio Sebastido after the large island CENTRAL! ATLANTIC COAST 161 between Rio and Santos. This flows eastward toward Cape Frio, for the name of which it is doubtless responsible, the current being considerably cooler than the average littoral water in these latitudes. There it meets and mingles with the main stream of the Brazil Current, which flows southward from the direc- tion of the Abrolhos Islands. The relatively low temperature of the Sdo Sebas- tido may be due in part to upwelling from intermediate depths close along the steep shore. At certain times of year, especially in midwinter (June-July), morning fogs hang above the adjacent coast, completely concealing the land, but these usually dissipate early in the day under the power of the rising sea breeze. The effect of the Sio Sebastiao is therefore, in miniature, like that of the Humboldt Current along the coast of Peru. Its meteorological and oceano- graphic consequences illustrate objectively the great influence of even very slight mean differences of temperature within the tropics. From the subjective angle, as pointed out by von Humboldt (1852, 1, 178) at the end of the eighteenth century, atmospheric variations of only six or seven Centigrade degrees in tropical South America are sufficient to produce in human beings the opposite sensations of heat and cold. The vicinity of Rio marks the beginning of a far more rugged type of coast than is to be found to northward. Here the eastern scarp of the Brazilian horst closely parallels the sea. Steep, rounded, rocky hills, of the type known as ilsenbergs, rise from the narrow lowlands. There are numerous sharply incised bays and, for the first time since leaving Trinidad, lofty islands represent sepa- rated portions of the continent. Ilha Grande (1000 meters), nearest to Rio, still has a relatively scanty human population. Sao Sebastido (1339 meters) is as high as the adjacent mainland shore, and as thickly populated. Santa Catharina, 48 kilometers in length and 600 meters in altitude, is the site of the capital city of the State of Santa Catharina, and has a population of close to 100,000. Santo Amaro, Ilha do Mar, Cananéa, and Ilha do Cardozo are other members of this chain of high continental islands, which are forested except where the land has been cleared for cultivation. In this respect they entirely resemble the mainland terrain and present a marked contrast with lower islands a little farther offshore. The latter, having no opportunity to benefit by the adiabatic cooling made possible by long and steep slopes, suffer from the charac- teristic ‘trade wind drought,”’ which has been mentioned so often before. For instance, the Alcatrazes Islets, about 24 kilometers southwest of Sao Sebastido, and at a considerably greater distance from any point of the Brazilian shore, are exceedingly barren even though the highest member of the cluster attains an altitude of 268 meters. The islets of Castilho and Figueira, to southward of Santos, are still better examples of arid spots not far from shore but too low to be effective in the local condensation of water vapor. Amaral (1921, 41) has given a description of a number of the small islands in this region, including particularly Queimada Grande, which lies about 64 kilometers southwest of the Bay of Santos and has an area approximating one square kilometer. Al- though in part well covered with small trees, Queimada Grande has a generally parched appearance and receives relatively little moisture when compared with 162 OCEANIC BIRDS OF SOUTH AMERICA the opposite mainland. It is surrounded on all sides. by great rocks, and is quite inaccessible except when the sea is calm. It is inhabited by a number of species of small land birds and by an extraordinarily large number of bird- eating snakes, the curious habits of which Amaral has described in detail. The resident sea birds of the region are identified by this author only as ‘‘mergul- hdes’’ and *‘gaivotas,’’ but one of his plates fortunately shows a group of what appear to be Brown Boobies, perched upon a large boulder. Doubtless the neighboring Alcatrazes Islets take their name from this same species. In most parts of Hispanic America pelicans are the first choice for the name alcatraz; wherever pelicans are lacking, the word is applied to boobies. Bryce (1912, 371) has called the scenery of the richly wooded eastern side of the Brazilian plateau, where it breaks down steeply towards the Atlantic, as beautiful as can be found anywhere in the equatorial regions. He had ap- proached this coast from the south, and when his steamer, . rounding a lofty cape, turned her prow shoreward to enter the harbour of Santos, . . . here at last were the tropics. Here was the region of abundant and luxuriant vegetation, a soft, moist air and a sea of vivid blue, with the strange thin-bodied, long-winged frigate birds hovering above it. As we came near enough to see the waves foaming on the rocks, an amphitheatre of mountains was disclosed, surrounding the broad, flat valley through which a river descends to form the port of Santos. To the north there ran along the coast a line of lofty promontories against which the surges rose. The mountains behind, all densely wooded, were shrouded with heavy mists, but the sun bathed in light the banks of the river, covered with low trees and flowering shrubs, and the gaily painted houses of the suburb which stretches out from the town of Santos, embowered in palm groves, to the white sands of the ocean beach. The coastal region is here low and narrow, and is sharply separated from the high interior by the Serra de Paranapiacaba, as it is farther northward by the Serra do Mar. The land is once again crossed by many rivers, all of which are characterized by an extremely rich aquatic life. In addition to the abundant resident birds native to the shore and the country behind it, there are also many snipes and plovers which migrate from the northern hemisphere, as well as visitors from southern South America, such as the Black-necked Swan, which comes to S4o Paulo to escape the rigors of the Patagonian winter season (Koenigs- wald, 1896, 332). The drowned character of the coast line increases southward. At Cananéa, for example, the coast is slit by narrow, branched straits, more or less parallel with the shore, which widen out to southward to form the inland sea known as the Mar de Tarapande. The Ilha do Cardozo and the Ilha Comprida lie on the ocean side of these waterways, which extend as sheltered natural canals all the way up the coast to Iguapé. Off the main entrance into Cananéa lies the small islet of Bom Abrigo, the welcome sight and name of which guide vessels through the channel between dangerous bars. Even more submerged is the appearance of the Bay of Paranagua, which cuts far into the state of Parana with numerous fiord-like arms, and makes a practicable approach to a broad valley rich in tropical vegetation. Two large islands in the mouth of the bay furnish shelter from the prevailingly strong surf and currents but, after entering, one finds a quiet, land-locked harbor, with steep-sided inner coves extending in CENTRAL ATLANTIC COAST 163 several directions (‘'N,’’ 1860, 327). Still farther southward, as at Pelotas, the Tertiary rock of the plateau is buried under a hundred meters of recent gravels and clays. From Cape Santa Martha Grande, south of Santa Catharina Island, the coast trends southwestward for nearly 500 kilometers, or as far as Rio Grande do Sul, in the form of long sandy beaches or ‘‘praias.’’ The mountains are here far inland and, since they draw away from the ocean still farther toward the south, where great lagoons lie behind the beach barriers, the shore is prevailingly arid for the same reasons noted as applying to low islands. Beyond the white beaches are dunes and flat sandy hills, with a scanty cover of stunted vegetation. The largest of the many fresh lagoons, lying in faults of the coastal terraces, is the Lagoa dos Patos. Does this body of water take its name from true ‘‘ducks,”’ or from the vast numbers of cormorants (Phalacrocorax olivaceus) that frequent it? Grafted on to its northern end is the deep bay of Porto Alegre, which is itself the drowned estuary of the Rio Jacuhy. The flood-head of the river has been sufficient for the Lagoa to maintain an egress through the barrier beach, in the form of the Rio Grande, which is a strait between lake and ocean rather than a river. The prevailing winds along this stretch of coast are northeasterly, but the southwesterly pamperos, which blow with great violence, are not uncommon during the winter, their effect sometimes reaching northward to Santos. At the latter city the annual rainfall] amounts to 2083 millimeters, with the dry season lasting only from June to August. A little to southward we meet the region of even rainfall régime, which extends down the coast beyond the es- tuary of the Rio de la Plata. Thus at Porto Alegre and Pelotas the rains are almost equally divided throughout the seasons. The mean annual temperature naturally lowers gradually from north to south in the coastal region. At Santos it is 21.9° C., at Porto Alegre 19.1°. However, in drawing away from the tropic of Capricorn the summer temperatures are reduced only slightly, those of winter much more markedly. Thus at Santos the means for the warmest and coldest months, respectively, are 25.5° C. and 19.1°; for Porto Alegre the corresponding figures are 24.6° and 13.6°. The cultivation of characteristic tropical culture plants, such as sugar-cane and coconuts, can be carried on along the coast about as far south as Porto Alegre (30° S.). As regards native seashore vegetation, we reach in the vicinity of Santos the southern limit of mangrove-growth along the east coast of South America. These plants, which have been mentioned so often in the account, comprise several species, of which the most important are the true mangrove (Rhizophora manglier) and the black mangrove or courida (Avicennia nitida). The first of the two grows mainly on flats overflowed by tidal rivers, while the second fringes more or less open seacoasts wherever the topography is favorable. The chief factors determining the distribution of the plants seem to be sufficiently saline water and a minimum annual temperature of about 10° C. That they are an excellent index of average climatic conditions is evidenced by their very dissimilar distribution on the easterly and westerly sides of South America. 164 OCEANIC BIRDS OF SOUTH AMERICA On the Pacific coast they extend southward only as far as the Peruvian border of the Gulf of Guayaquil (about latitude 3° 30’S.). At this point the absence of estuaries, together with climatic conditions associated with the Humboldt Current, impose an abrupt bar to their extension. Along the Caribbean and Atlantic coasts they are characteristic, as we have seen, of suitable localities as far southward as latitude 24° S. or a little beyond. It is rather curious that whereas the southward growth of reef corals falls short of that of the man- groves on the east coast of South America, the same organisms have contrived to find at least a feeble development for a long distance south of any mangrove- growth on the western side of the continent. In general the range of mangroves may be said to be substantially similar to that of tropical jungle vegetation in well-watered low country immediately behind the shore. It is not very different, moreover, from the present range of feral coconut palms along the coast. The latter plant is, of course, strongly halophytic. It attains its greatest vigor and luxuriance at the seashore, and is usually the first tree to become established upon newly exposed tropical beaches and coral reefs. Humboldt (1852, 290) reports that when these palms were planted in the inland missions of the Orinoco, it was customary to put half a bushel of salt into the hole receiving each nut. The several organisms, both plant and animal, just mentioned as examples of tropical distribution are confined to either the land or the sea side of an attenuated geographic zone of torrid coastal lowlands. With only brief inter- ruptions, this extends for more than 11,000 kilometers along the borders of South America, or from near the southern boundary of Peru northward and eastward around the continent, and then southward to about latitude 27° S., or to the neighborhood of the island of Santa Catharina, Brazil. In most places this coastal belt is very narrow; only in the region of the Amazon estuary does it extend inland as much as 500 kilometers. From southern Brazil around the southerly tip of South America, and thence up the west coast to the northern end of Chile, such flat or undulating coastal country as has just been referred to is replaced nearly everywhere by mountainous regions, or by plateaus of con- siderable height which are steep along the edges facing the sea so that the in- land limits of the coastal strip are at least very clearly marked. The two types of coastal topography, in combination with the distinct character of the littoral water found along each region, correspond in the main with the grand division between the pan-tropical and the sub-antarctic realms of ocean and shore life. Latitudinally the two parts are highly asymmetrical, each tending to center about its respective pole of an axis that crosses the con- tinent of South America from northeast to southwest. Allowing for certain variations in ranges due to local conditions, particularly to Humboldt Current phenomena on the west coast, we may say that the northeasterly moiety of the South American coastal outline, from southern Brazil to Peru, lies within the range of reef corals and mangroves; of the Portuguese man-o’-war (Physalia), flying fish, dolphins (Coryphaena), and other tropical pelagic creatures; of frigate-birds, tropic-birds, and equatorial types of boobies. Its beaches, more- CENTRAL ATLANTIC COAST 165 ~ =, == SS — Fic. 23. Complementary examples of oceanic distribution in South America, as correlated with surface temperatures. Coarse shading shows the range of a pan-tropical man-o’-war bird (Fregata magnificens), fine shading that of a sub-antarctic tern (Sterna hirundinacea). The heavy, transoceanic lines mark the approximate mean positions of the Antarctic and the Sub-Tropical Convergences, the course of the latter being very uncertain in the Pacific. / over, are the breeding sites of sea turtles. On the other hand, the southwesterly part of the continental circumference, from southern Brazil to Peru by way of Cape Horn, comprises the totally different and supplementary zone of the giant kelp (Macrocystis); of sub-antarctic seals and sea-lions; of penguins, white- breasted cormorants, Wandering Albatrosses, Cape Pigeons (Daption), and other birds which are at home in cold seas. The respective assemblages of life should be thought of in relation to the oceanic zones discussed above in the section on hydrology. In the northeasterly, or generally tropical, section of the South American coast, many kinds of inland marsh birds, some of which have already been given passing mention, reach the ocean shore. Other species, such as the West Indian Flamingo and a number of Limicolae, are almost wholly confined to beaches and flats washed by salt water. Examples of both of these aggregations 166 OCEANIC BIRDS OF SOUTH AMERICA of tropical or sub-tropical birds, which can by no means be called oceanic but which are useful, nevertheless, as indicators of oceanic climatic influences, are listed below. In some instances the coastal range is typical or exclusive; in others it is merely a coastal contact dependent upon mangrove-growth; in still others, the species concerned inhabit treeless swampland, or even beaches lying in front of arid plains. The assigned limits of coastal range are to be regarded as only approximate. Probably in no instance, moreover, is the littoral distribu- tion continuous throughout the range indicated; rather, its pattern is deter- mined by various local ecological factors. In the following list, subspecific discrimination of the birds is omitted. Water-TurRKEYs Anhinga anhinga. Ecuador to southern Brazil. HERONS Ardea cocoi. Peru to Argentina. Butorides striatus. Colombia to Uruguay. Florida caerulea. Peru to Argentina. Leucophoyx thula. Northern Chile to Uruguay. Hydranassa tricolor. Ecuador to eastern Brazil. Nyctanassa violacea. Peru to southern Brazil. Botaurus pinnatus. Colombia to southern Brazil. Woop InisEs Mycteria americana. Peru to Argentina. IBIsEs Mesembrinibis cayennensis. Ecuador to Argentina. Guara rubra. Ecuador to southern Brazil. SPOONBILLS Ajaia ajaja. Peru to Argentina. FLAMINGOES Phoenicopterus ruber. Galapagos Islands to Guiana. RaILs Rallus longirostris. Peru to southern Brazil. Aramides cajanea. Peru to Uruguay. PLOovERS Charadrius collaris. Peru to Argentina. Somewhat similar to the coastal distribution of the foregoing sixteen species is that of the two most characteristic terns of South American tropical rivers. These are Phaetusa simplex and Sterna superciliaris, a large and a small species, respectively, both likely to be found along almost any part of the coast between eastern Colombia and Uruguay or Argentina. These two terns, like the marsh CENTRAL ATLANTIC COAST 167 birds named above, are to be regarded as eminently continental rather than oceanic. Since they belong to the fauna of great rivers, they are confined almost altogether to the slopes of Caribbean and Atlantic drainage, and have not occupied the narrow western side of the Andean watershed. The Large-billed Tern (Phaetusa), however, has been at least once or twice recorded from western Ecuador, where the River Guayas alone simulates a fluvial system worthy of comparison with those that bear water toward the Atlantic. In the estuaries of the Caribbean and Atlantic rivers, and along the shores between, both of these terns reach salt water. The Red-billed Tern (Thalasseus eurygnatha) has likewise practically the same coastal range, but this species is typically littoral rather than fluviatile. It inhabits many of the Caribbean islands, and others along the Atlantic border of the continent. Let us for the moment leave the pan-tropical birds, whether of continental or pelagic affinities, and consider the waterfowl of the Atlantic west wind belt. We may say that Cape Frio, washed by the cool Sao Sebastido, represents at certain seasons of the year the threshold of the south. It is along this part of the Brazilian coast that we begin to encounter sea birds from the pan-antarctic zone in watery territory more properly belonging to tropical species and to widely distributed continental water birds. Perhaps the cool coastal current has much to do with the regular occurrence of southern species in the bight of the coast that stretches from Cape Frio to Cape Santa Martha Grande. At any rate, here the breeding ranges of a southern tern (Sterna hirundinacea) and of the Kelp Gull (Larus dominicanus) overlap that of the Brown Booby. Penguins, Cape Pigeons (Daption), and an occasional albatross, moreover, visit waters within the regular beats of the man-o’-war bird and of tropical terns. The von Iherings (1898, 113; 1904, 345; 1907, 1) have discussed the birds of Sado Paulo and the states to southward, and Luederwaldt (1929, 1) those of the island of Sao Sebastiéo. The following sea birds of southern Atlantic affinities are reported: Spheniscus magellanicus. Frequently caught in the nets of fishermen; often dies in great numbers during midwinter (August) along the beaches between Rio Grande do Sul and Paranagua. Reported from Ilha de Sao Sebastiao, Santos, Iguapé, and Guaruja. Diomedea melanophris. Rio de Janeiro, Santos. Also an albatross, listed by Luederwaldt as ‘‘Thalassogeron sp.,"’ has been found dead on the beach of the Ilha de Sao Sebastiao. Daption capensis. Ilha de Sio Sebastiao, Iguapé, Rio Grande do Sul. Spix and Martius (1824, 1, 119), moreover, report Daption capensis and Phaéthon aethereus seen together in latitude 18° S., longitude 35° 20’ W., close to the Abrolhos Islands. There could hardly be a better example than this of the meeting of extremes! _ Pachyptila. Whale-birds, under several different names, have been recorded from Ilha de Sao Sebastiao, Iguapé, Santos, and even from Bahia. The specific determinations, however, do not inspire confidence. 168 OCEANIC BIRDS OF SOUTH AMERICA Procellaria aequinoctialis. Iguapé, Bahia. Pterodroma macroptera. Dead on the shores of Ilha de Sado Sebastiao. Oceanites oceanicus. Many points along the coast of Brazil. Von Ihering and Luederwaldt furthermore list the Manx Shearwater (Puffinus puffinus) as a migrant petrel from the northern hemisphere which has been collected along the Sio Paulo coast. Among resident and breeding species of water birds they name the following: Sula leucogaster. Phalacrocorax olivaceus. ‘‘Lives equally on the coast, in lagoons, and along rivers of the interior’’ CH. von Ihering). Fregata magnificens. (Listed by these authors under the names aquila and minor. All specimens I have seen from southern Brazil, however; prove to be magnificens ). Larus dominicanus. Phaetusa simplex. Sterna hirundinacea. Sterna superciliaris. Sterna trudeaut. Thalasseus eurygnatha. Rynchops nigra intercedens. The skimmer and several of the terns are, of course, chiefly birds of the bays, estuaries, and lagoons. The mixture of tropical and southern species is a further indication of the overlapping avifaunas of southern Brazil. Von Ihering states that tropic-birds (Phaéthon) are not found along the coast of Sao Paulo. This is interesting in view of the presence of both the Atlantic species as breeding birds at Fernando Noronha and Ascension. Since neither Sula dactylatra nor Sula sula is mentioned as occurring in these waters, it is highly probable that these two distinctly tropical boobies, as well as the tropic-birds, do not pass to southward of the clear and warm littoral ocean that ends somewhere in the vicinity of Cape Sao Thome, or not far from the tropic of Capricorn. Hentschel (1933, 121) speaks of the tropic-birds as reaching a maximum limit of *‘40° S. latitude’ in the South Atlantic. The statement, however, is probably a slip of the pen, for the same author's accompanying distributional chart (pl. VID) shows that during the cruise of the ‘Meteor’ no field observations of tropic- birds were made even as far south as latitude 20° S. Southward from Rio Grande do Sul, the Brazilian coast continues of the same general character as described above, all the way to the Uruguayan border. A little north of Point Palmar, which is the first of several salients that turn progressively toward the estuary of the Rio de la Plata, are the Coronilla Islands, where there are seal rookeries under the jurisdiction of the Uruguayan govern- ment. Smith, who has investigated these and neighboring islets, writes (1927, 287): The most northern islands on the coast of Uruguay are the Coronilla group, lying in latitude 33° 56’ S., about 17 kilometers to the southward of the Chuy River which marks the boundary CENTRAL ATLANTIC COAST 169 between Uruguay and Brazil. These islands are resorted to by both fur seals and sea-lions, and are the northern limit of the range of these animals on the east coast of South America. The group consists of two main islands and several islets, and lies about 214 kilometers offshore, off a prom- ontory sometimes designated as Punta de los Loberos. With reference to the northernmost breeding stations of southern pinnepeds in South America, it is worth noting that “‘lobo’’ is the Brazilian, as well as the Spanish-American, word for sea-lion, and that there are two small islets called Lobos—Grande and Pequena, respectively—at about latitude 26° 30’ S., just south of the Ilha de Sio Francisco. Perhaps in former times the sea-lions or fur seals, or both, ranged somewhat farther toward the edge of the tropics than their reduced population now extends. South American seals of all sorts graphically reflect by their distribution the same kind of oceanographic phenomena which we have attempted to inter- pret through the distribution of sea birds and of many other marine and neritic organisms. The southern sea-lion (Otaria byronia), for example, ranges from these small islands, just north of the estuary of the Rio de la Plata, southward along the Patagonian coast to Cape Horn. It reaches the Falklands, but not South Georgia; neither does it cross Drake Strait into truly antarctic waters. From islets off Tierra del Fuego the range stretches continuously northward along the western coast of the continent as far as the Humboldt Current is in contact with the shores. The northernmost ‘‘loberias,’’ or sea-lion rookeries, are in the neighborhood of the westernmost projection of South America, at Punta Parifias, Peru (4° 45’ S.). Before the destructive advent of white men, the so-called antarctic fur seal CArctocephalus) and the sea-elephant (Miérounga) each had a very similar distri- bution. The former animal has now been largely decimated, and the sea- elephant wholly exterminated, within the South American field. Just as the three species of seals thus far mentioned marked the cold waters of high southern latitudes, or zones of cool upwelling within the tropics, so the nearly extinct West Indian seal (Monachus) was once a symbol of the warm Caribbean. Inci- dentally, the southeastward extension of the West Indian seal seems to have been limited by the same physical attributes of the ocean water which have prevented the Brown Pelican from invading the northeasterly segment of the South American coast. I can see no other reason, at any rate, why both pelican and seal should not have reached or rounded Cape Sao Roque. Off Cape Polonia, Uruguay, which is a conspicuous rectangular promontory flanked by broad sandy beaches that merge into high dunes, are the Castillos and Torres groups of seal islands. Farther westward, near the famous ‘‘East Point’’ of the estuary, are the larger and more important Lobos Islands. The Lobos Islands, lying 10 kilometers off Punta del Este, Department of Maldonado, are usually cited in works of reference as located in the mouth of the Rio de Ja Plata. The trend of the coast from Montevideo to Maldonado is in general a continuation of the left bank of the estuary of the Plata, but it does violence to the physical and biological facts to regard the Lobos Islands, with their typical marine littoral fauna and surrounded by water of high salinity, as being in the mouth of the Plata. The Lobos group consists of a main island, an outlying low islet, and a number of small de- 170 OCEANIC BIRDS OF SOUTH AMERICA tached rock masses over which the surf breaks. Lobos Island proper is about one kilometer long and half a kilometer wide, with a maximum height of 20 meters, and is composed mainly of exceed- ingly rugged granite ledges, boulders, and cliffs, in places left in fantastic shapes by the volcanic upheaval of which evidences remain in small smooth laval areas and outcropping honeycombed laval rocks in the interior of the island. A shallow soil of sand, shells, and humus supports a sparse stunted vegetation of which conspicuous elements are a harsh brake and a diminutive cactus. Sandy beaches are few in number and of limited extent, the principal beach, at the northwest end, being the only safe boat landing and having determined the location of government houses con- nected with the sealing industry, a lighthouse, and a wireless station. There are no harbors, and landing is at times difficult, if not precarious, owing to the heavy swell and strong currents. The southern part of the island has the most elevated and broken shore, a striking feature being a huge, semi-detached castle-like buttress, with precipitous sides and numerous ledges and crevices to which the seals resort, the base being surf-beaten (Smith, 1927, 279). Smith notes, furthermore, that the waters about all of these essentially temperate islands are infested with ravenous sharks of sufficiently large size to prey upon the seals. His remarks upon the resident bird life, in the following description of the main island of the Coronilla group, is particularly important for our purposes (287): The larger island is landward and eight-tenths of a kilometer long. The low northern end is composed of granite ledges and boulders, with a considerable soil-covered area planted with cane and oats about an abandoned sealing plant. The higher southern end is very rugged, with extensive fissured granite ledges and huge boulders. The shore has no landing places but inside the surf- washed boulders there are half a dozen flat sandy or shelly areas. In some places the shells, washed in by the storm waves, are in great variety and of exquisite beauty and are in beds several meters thick. Wherever there is vegetation and in and about the abandoned buildings, there exist untold thousands of large land snails, including two edible species introduced into Uruguay from Europe many years ago and now found all over the country. Conspicuous birds resident on the island were a black-backed gull, abundant and nesting, a few eggs observed, many downy young running about; a small tern, nesting; the teru-tero, or lapwing, a number of old birds observed and several young caught by the Uruguayan sailors who stated that the birds are easily tamed and make interesting household pets; the black oyster-catcher, several adults seen and a nearly grown bird and two downy young brought aboard the vessel; a black cormorant, numerous; a red-breasted grebe, a nearly full-grown young of the year caught in tall grass; and a colony of orioles nesting in the cane-brake. The gull mentioned is Larus dominicanus; the tern probably Sterna hirundin- acea; the lapwing Belonopterus chilensis lampronotus; the black oyster-catcher Haematopus ater, here doubtless at the extreme northern limit of its breeding range on the Atlantic coast; the cormorant Phalacrocorax olivaceus, with which we have already become acquainted in a wide variety of climatic and edaphic environments; the grebe, presumably Colymbus chilensis. Also, Wetmore (1926, 134) tells us that the Brown-hooded Gull (Larus maculipennis) nests in abundance on rocky islets along the coast of the Department of Rocha, Uruguay, which can mean only the Coronillas, the Lobos, or neighboring groups. The list of residents has a prevailingly southern stamp, which is heightened by Smith's reference to the annual appearance in large numbers of penguins about the northern seal islands of Uruguay (/. c., 292). Every winter (July- August), he writes, penguins resort by thousands to the islands and to the extensive sandy beaches at Cape Polonia. Then, as the season advances and the CENTRAL ATLANTIC COAST 171 air and water become warmer, they sicken and die. ‘‘Not one penguin survives until the next fatal migration occurs.'’ As many as 3000 have been known to succumb at Cape Polonia during a single season. From the skins of many the local sealers make up various articles of wearing apparel. Smith is certainly incorrect in identifying these penguins as Pygoscelis papua. They are undoubtedly the Magellanic Penguins (Spheniscus magellanicus) under- taking the normal winter exodus from nesting grounds along the Argentine coast, as described hereafter in the biography of the species. 5. Tue ATLANTIC SuB-TRopIcAL IsLANDs Leaving the continent once more, we may now visit several oceanic islands which I have already designated as sub-tropical both because of their climatic position and the constituency of their bird life. South Trinidad, with the neighboring Martin Vas group, and St. Helena comprise the entire list of these. St. Helena has, of course, even less claim than Ascension to be called a South American island; nevertheless, it is a link in the distributional chain of South Atlantic sea birds. While Tristan da Cunha lies, as heretofore stated, on the convergence between sub-tropical and sub-antarctic waters, its zodlogical associations are mostly with the latter zone, despite the presence of breeding noddies. The Tristan group may therefore best be discussed in conjunction with Gough Island, which, like the Falklands, is thoroughly sub-antarctic. a. South Trinidad and Martin Vas. South Trinidad or, in the Portuguese spelling, Trindade, is 1207 kilometers from the continent, with dimensions commonly recorded as about 6.4 by 3 kilometers, and an altitude of about 600 meters. Although geologically rather similar to Fernando Noronha, having peaks of phonolite associated with basaltic lavas, Trindade differs from the less lofty equatorial island in that it nearly everywhere rises steeply from the ocean; there is almost no strand, and very little coastal fringe of any other sort. Coral reefs are mentioned by Sim- mons (1927, 28) but by none of the earlier visitors; their presence needs con- firmation. The geographic position of the island is toward the southerly edge ot the trade wind zone. Rainfall is sufficient to support a wealth of tree ferns and other evergreen vegetation on the upper windward slopes, but the northerly sides of the mountains are barren. Owing to the porosity of the volcanic soil and the absence of stratified rocks, most of the precipitation is absorbed beyond redemption; the brooks and waterfalls are transient even if torrential, and the few springs are untrustworthy. One stream, flowing through a cluster of ruined stone huts on the northern shore, may perhaps be perennial. Trindade has more than once been occupied by resident human beings, but can hardly be said ever to have supported a permanent settlement. However, Rockwell (1932, 434) infers that it has been again used within the last decade as a place of detention for Brazilian political prisoners. So much that is romantic and eerie is associated with the history and appear- 172 OCEANIC BIRDS OF SOUTH AMERICA SOUTH TRINIDAD ISLAND Crista de Gallo| Pt Scale of miles A ! Kilometers 4 i Heights in meters Vallado Pt + Ninepin (Maury 1. Cachoeira \Bay Five Farilt -s Pt 2 | Natural Arch MARTIN VAS ISLANDS ? 4 1 MILE ° Yn I KILOMETER Fic. 24. ance of Trindade that it would be unfair, even in a scientific discussion, to give no suggestion of this. The following paragraph is from a report made after my own youthful visit to the island in April, 1913. In the paper referred to, incidentally, I have cited a bibliography of 19 titles, dealing with the principal visits of naturalists to Trindade up to that date. East of the coast of Espirito Santo some seven hundred miles lies a fairy island. Alone in the tropical ocean, piled up in peaks as fantastic as tossing waves, and overhung with pennons of torn clouds which seem to flutter from the summits, Trinidad has exercised a strange charm upon the imaginations of all who have but seen its silhouette on the borderline of sky and sea. During four centuries it has been a landmark in the trade routes of the South Atlantic, often sought by sailing vessels as a check upon their nautical reckonings. Before the days of steamers it was a veritable signpost at a crossroads of the sea, yet few indeed are the travelers who have set foot upon its crumbling shore. Pirates in the old times, whalers, treasure-seeking adventurers, ill- fated colonists, in their turn have come to Trinidad and gone; the island seems unfalteringly to forbid the encroachment of permanent habitation. None who have felt its presence can speak or think of it unstirred; even the prosaic pages of the ‘South Atlantic Pilot’ become alluring at the account of Trinidad, and the Director of the British Antarctic Expedition of 1901, though he surveyed the islet with the critical eye of science, was deeply impressed by “‘the dream-like appearance of this remarkable cluster of volcanic peaks in the early tropical dawn’’ (Murphy, 1915, 332). All over Trindade the brittle, standing, residual rock structure has assumed grotesque forms through extreme weathering. The summits of the ridges are a succession of needle-like pinnacles. The island has given many visitors the impression that through erosion and landslides it is fast crumbling to bits. Needless to say, the crossing of the mountainous ridge, or an ascent to the highest point, is a relatively formidable undertaking, although both have been ATLANTIC SUB-TROPICAL ISLANDS 173 many times accomplished. During my own single day on the waters about Trindade, we found it impossible to land because of the heavy swell. Twice since that date, however, the island has been exhaustively investigated, first by a Brazilian scientific expedition (Ribeiro, 1919, 171), and later by the ‘Blossom’ party of the Cleveland Museum of Natural History (Simmons, 1927, 28), the members of which spent nearly six weeks at the island during 1925. For a month of this period Messrs. R. H. Rockwell and Alan Moses dwelt in a cave on a hillside, just out of reach of the furious surf, while collecting and pre- paring a comprehensive series of the fearless resident birds (Rockwell, 1932, 435). In the destiny of its vegetation, Trindade may be going the way of Ascension, although its current stage is several centuries later than that of the latter island. When the astronomer Halley landed at Trindade in April, 1700, its windward coast was well covered with forests of large Brazilwood trees (Caesalpinia). These trees subsequently all died, the last mention of living examples in the groves harking back to the eighteenth century. There has been a good deal of speculation as to the cause of such destruction, the explanations offered includ- ing an emanation of volcanic gases, a sharp decline in the amount of rainfall, and the activities of goats liberated by Halley (Copeland, 1882, 269). In view of the history of St. Helena forests, discussed below, it is likely that the goats offer an adequate explanation for the destruction of the trees of Trindade. The creatures would first feed upon the undergrowth, while increasing enormously in numbers, and would take next to the leaves and bark of the trees. As food became used up, the goat population would in turn decline, until a natural balance had been attained. The goats of Trindade, long thought to have died out completely with the passing of the Brazilwood trees, have been observed, together with hogs, on the secluded upper slopes of the island as recently as 1925 (Simmons, 1927, 28; Rockwell, 1932, 436). Simmons reports that they still appear to subsist entirely upon vegetation, while the feral hogs feed chiefly upon land crabs. The novelist Captain Marryat includes a picturesque and factual account of the dead forests of Trindade in his first work of fiction, “‘Frank Mildmay, or the Naval Officer,’’ which was published in 1829. His words are as follows: Here a wonderful and most melancholy phenomenon arrested our attention. Thousands and thousands of trees covered the valley, each of them about thirty feet high; but every tree was dead, and extended its leafless boughs to another—a forest of desolation, as if nature had at some particular moment ceased to vegetate! There was no underwood or grass. On the lowest of the dead boughs, the gannets, and other sea-birds, had built their nests in numbers uncountable. Their tameness, as Cowper says, ‘was shocking to me.’ So unaccustomed did they seem to man that the mothers, brooding over their young, only opened their beaks in a menacing attitude at us, as we passed by them. How to account satisfactorily for the simultaneous destruction of this vast forest of trees was very difficult: there was no want of rich earth for nourishment of the roots. The most probable cause appeared to me, a sudden and continued eruption of sulphuric effluvia from the volcano; or else, by some unusually heavy gale of wind or hurricane, the trees had been drenched with salt water to their roots. During the voyage of the ‘Terror,, McCormick (1884, 23) landed twice on 174 OCEANIC BIRDS OF SOUTH AMERICA Trindade in January, 1840. At that date the upper slopes were still covered with great numbers of dead trees, barkless and blanched, mostly fallen and scattered about in wild confusion, but with here and there one standing erect in the soil. After the beginning of the present century, Knight (1907) added the following account, which describes the contemporary status of the Brazil- wood forests: The mountain slopes were thickly covered with dead wood—wood, too, that had evidently long since been dead; some of these leafless trunks were prostrate, some still stood up as they had grown. . . . When we afterwards discovered that over the whole of this extensive island, from the beach up to the summit of the highest mountain—at the bottom and on the slopes of every now barren ravine, on whose loose-rolling stones no vegetation could possibly take root—these dead trees were strewed as closely as it is possible for trees to grow; and when we further perceived that they all seemed to have died at one and the same time, as if plague-struck, and that no single live specimen, young or old, was to be found anywhere—our amazement was increased. . Looking at the rotten, broken-up condition of the rock, and the nature of the soil, where there is a soil—a loose powder, not consolidated like earth, but having the appearance of fallen volcanic ash—I could not help imagining that some great eruption had brought about all this desolation; . . . . I think this theory a more probable one than that of a long drought, a not very likely contingency in this rather rainy region. The very recent notes and photographs of the ‘Blossom’ party give a similar picture, except that still more of the dead boles have fallen, while the branches and the prostrate logs have progressed several stages nearer final dissolution. The condition of this rotting vegetation has a very important effect upon the history and future of the resident sea birds, a subject to which I shall return below. Simmons collected about thirty species of plants on Trindade, making a considerable addition to the previously known flora. The tree fern, so con- spicuous on the plateaus and higher slopes, is an endemic species (Cyathea Copelandi). The lower limit of its growth was determined by the naturalists of the ‘Discovery’ to be at an altitude of about 335 meters. In addition to the vascular plants, there are a number of species of mosses and lichens, including a tree-infesting Usnea. Before discussing the bird life of Trindade, it will be well to set down a brief description of the neighboring Martin Vas Islets, which lie about 50 kilo- meters to eastward. The altitude of the highest of these has been variously given as from 90 meters to twice that. At any rate, the cluster is high enough to be visible from the shore of Trindade. There are three main islets, lying along a north-south line, the two northern ones close together, the southern member about 4 kilometers from the other end of the chain. All are relatively inacces- sible, the middle and tallest island, in particular, towering abruptly from the sea toward its flat plateau, which is covered with luxuriant grass or sedge. Round about are many fantastic rocks, some in the form of spires, others domed and arched. The geological composition is evidently similar to that of Trindade. The Martin Vas group was briefly visited by a whaleboat party from the schooner ‘Blossom’ in December, 1924 (Simmons, 1927, 27). The bird life of Trindade and Martin Vas differs from that of the Atlantic ATLANTIC SUB-TROPICAL ISLANDS 175 islands to northward in the apparent absence of the two large intertropical boobies (Sula dactylatra and S. leucogaster). Positive and more significant differ- ences are found in the presence of the Lesser Frigate-bird (Fregata ariel), and also of a species of Greater Frigate-bird (Fregata minor); we have not previously encountered either of these during our circumnavigation of South America. Still more important among the new elements is a petrel (Pterodroma armin- joniana) which is a member of a group widely distributed in the Sub-Tropical Zone of the southern hemisphere. The fact that neither of the Atlantic tropic- birds (Phaéthon) has yet been reported from Trindade or Martin Vas need not be emphasized until it becomes certain that no member of the genus is to be found there. The islands have thus far been investigated at haphazard seasons of the year, and it is worth noting that for a long period both the Noddy and Sooty Tern were overlooked. There seems to be no longer any doubt, however, that the islands do not harbor a single species of native land bird. The resident avifauna, as known to date, comprises the following eight species: Pterodroma arminjoniana. A petrel of several puzzling plumage phases, no less than four of which have been described under as many specific names. It is related to P. neglecta of the sub-tropical South Pacific and may be regarded, indeed, as the Atlantic representative of that species. Like its close relatives, the Trindade petrel is a surface-nesting bird, laying its eggs in hollows and niches of the rock. It inhabits both Trindade and the Martin Vas Islets in vast numbers. Except for Oceanodroma castro at St. Helena and Ascension, P. arminjoniana is the only petrel certainly known to breed in any part of the Atlantic between Tristan da Cunha and the West Indies. Sula sula. An abundant species which appears to be the only resident booby of Trindade. It is a bird of critical interest because of its dependence upon vegetation for its nest and nesting sites. Everywhere in the world, so far as I have been able to determine, this booby constructs its nest of twigs and places it among the branches of shrubs or trees, upon the fronds of palms, or upon stumps or fallen logs. Curiously enough, it lays but one egg, as do the frigate-birds and fairy terns, whereas the ground- nesting boobies usually lay a complement of two or more eggs. The fairy terns, however, are content to nest upon either wood or rock, and the frigate-birds can at least be forced by circumstances to abandon their shrub-nesting habit and to build upon ledges or even upon bare ground close to the edge of a plateau above the sea. But in the case of the Red-footed Booby,'I have thus far found but a single instance (reported upon in the biography of the species) of a nest that is not supported upon some sort of plant-growth, dead or alive. The dis- appearance of suitable vegetation from any island means, indeed, that Sula sula will cease to reside and breed upon it. No doubt the practical abandonment of Ascension Island by this species took place long ago with the dying off and eventual disintegration of the original trees. Trindade seems to be following the same course, except that there is a considerable amount of arborescent vegetation in addition to the blighted and vanishing groves of Caesalpinia. 176 OCEANIC BIRDS OF SOUTH AMERICA The ‘Blossom’ expedition photographs show the Red-footed Boobies of Trindade invariably nesting on weather-beaten and decaying wood, however small the fragments may be. Mr. Rockwell tells me, moreover, that during his visit of a month he never saw a nest, egg, or chick of the booby in any other situation. Most of the nests were of small size for this species, a result which might be expected in view of the difficulty the birds would have in finding a ready and abundant supply of small sticks, such as they ordinarily use. All of this leads one toward an interesting speculation regarding the future of the Red-footed Booby at Trindade. As the remnants of the Brazilwood become reduced to dust, will the boobies tend to confine themselves to cramped nesting sites in the bushy growths on the heights of the island, or will they make the seemingly simple yet unprecedented transition from wood-nesting to rock-nesting? Fortunately, the answer to this question can be learned within a measurable length of time, and a marked reduction in the booby population of the island would probably mean that the history of the species at Ascension _is by way of being repeated at Trindade. At the Martin Vas Rocks, where there is no woody growth, the Red-footed Booby does not nest. Fregata minor. At Trindade we find this large species of man-o’-war bird at its northernmost breeding station in the Atlantic. The statement would need modification if it should become known that the man-o’-war bird formerly residing at St. Helena belonged to the same species. In other oceans Fregata minor is very widely distributed, extending its range into the North Pacific as far as the Hawaiian group. As previously noted, the Trindade form apparently does not reach the Brazilian coast, where it is replaced by F. magnificens. Fregata ariel. The Lesser Frigate-bird is another new species in our list, find- ing at Trindade its sole breeding station in the Atlantic. Like Fregata minor, it appears to be highly sedentary. Both forms of frigate-bird were observed about the Martin Vas Rocks by the ‘Blossom’ party. Sterna fuscata. The Sooty Tern nests in December and January, its breeding areas being rather restricted. One colony occupies rocky platforms around tide- pools of the northern islet of the Martin Vas group, and a second certain islets known as the Tents, off Trindade. Anous stolidus. Abundant at both Trindade and Martin Vas. Anotis minutus. The Black Noddy inhabits Martin Vas but, so far as known, not Trindade. Nicoll (1906, 673) obtained one specimen. The ‘Blossom’ records refer to it as a species “‘breeding in nests plastered against an almost perpendicular cliff’’ at the main islet of Martin Vas (Rockwell, 1932, 434). Gygis alba. Abundant at both Trindade and Martin Vas. b. St. Helena. St. Helena is 2897 kilometers from South America and only 1835 kilometers from the nearest point of the African coast. It is larger than any of the oceanic islands we have yet considered, being 16.5 kilometers in length and nearly as ATLANTIC SUB-TROPICAL ISLANDS 177 ST. HELENA ISLAND Miles 2 3 Kilometers ie a ee 3 Heights in Meters Horse Pasture Pt Gill Pt «George I. Fie. 25. broad, with an area of about 122 square kilometers. It is wholly of volcanic origin, although the surface is much altered through erosion and redeposition, and on all sides except the southerly it has wave-cut cliffs ranging up to 600 meters in height. The curved mountainous ridge of the island is part of the rim of a great crater, characteristically breached toward the southerly or wind- ward edge. It culminates in Diana’s Peak, which is 824 meters above the sea. In the words of Darwin (1933, 409), the island looms like a huge castle from the ocean, with a wall built of successive streams of black lava around its whole bold coast. Near the shore the lava is entirely destitute of vegetation, but in the higher and central parts the rocks have decomposed into a clayey soil of many colors, which in the rainy seasons produces a singularly bright green pasture. At an altitude of 500 meters the modern vegetation possesses a decided British aspect, due to the fact that the introduced plants are tenfold more numer- ous than the indigenous, which number fewer than sixty. Of the thirty-eight species of native flowering plants, many of which have become extinct, all but one are endemic. At St. Helena the southeast trade wind blows nearly continuously, the con- 178 OCEANIC BIRDS OF SOUTH AMERICA densation causing a canopy of clouds to hang during much of the time at the level of the peaks. The condition accounts for a significant local expression concerning ‘‘covered days.’’ Strong trades, such as hold particularly during September and October, are the nearest approach to a storm that the island experiences, thundersqualls being all but unknown. As Meliss (1875, 387) ob- serves, it is remarkable that when the trade winds “‘blow with fury round the coast, even driving ships to sea from the snug anchorage on the leeward side, a perfect calm exists on the mountain ridge at Diana’s Peak.’’ The usual shallow- ness of the trade wind layer over the oceans is, indeed, a point that deserves stressing. The ‘Meteor’ party encountered days during which the whole moving body of atmosphere extended upwards not more than 50 meters above the surface of the Atlantic, with the thermometers recording 27° C. at the fore- masthead and only 19° on deck (Spiess, 1928, 302). St. Helena, like all steep, high islands, is a center of other curious vagaries of the winds, some of which doubtless have an important effect upon the selec- tion of nesting sites by sea birds. Darwin (1933, 413) writes: The only inconvenience I suffered in my walks, was from the impetuous winds. One day I noticed a curious fact; standing on the edge of a plain terminated by a great cliff of about a thou- sand feet elevation, I saw at the distance of a few yards, right to windward, some Tern struggling against a very strong breeze, whilst where I stood the air was quite calm. Approaching close to the brink I stretched out my arm, which immediately felt the full force of the wind. An invisible barrier of two yards wide, separated a strongly agitated from a perfectly calm air. The current meeting the bold face of the cliff must have been reflected upwards at a certain angle, beyond which there would be an eddy, or a calm. While long droughts occasionally occur at St. Helena, rainfall is usually moderately plentiful. There are both summer and winter peaks of precipitation, the summer rains being relatively heavy, but the winter rains (March to August) more continuous. The annual rainfall in the highlands amounts to 990 milli- meters, or about double that of Ascension, while the coast always feels more or less of the “‘trade wind drought.”’ Streams are very numerous, and there are said to be no less than 212 springs of fresh water. There are a few records of floods that have carried away trees, cottages, and stone walls, and have washed down surface soil in such quantity that the sea around the island looked like mud. Such phenomena, and even the normal run-off from the island, account for an enormous increment of terrigenous materials in the surrounding ocean waters, with a consequent direct and secondary enrichment of the neritic life (Hentschel, 1933, 165). The ‘Meteor’ naturalists were impressed with the abundance of both fish and terns at St. Helena (Spiess, 1928, 231). It is but natural that in so long-inhabited an island man and his creatures have wrought an overwhelmingly destructive effect upon the original biota. In 1683, St. Helena had a human population of 500; it has since risen to a maximum of between 4000 and 5000. It requires considerable faith on the part of the modern traveller to realize that this prevailingly barren island was once densely forested, but the testimony of Sir Joseph Hooker, as quoted by Meliss (1875, 221), leaves no possible doubt of it. When St. Helena was discovered, ATLANTIC SUB-TROPICAL ISLANDS 179 in 1502, it was entirely covered with forests, the gumwood, ebony, and other Mative trees even drooping over the tremendous precipices that overhang the sea. Hooker writes: The destruction of the Madeira forests . . . was by fire. A much more insidious agency has operated with tenfold greater effect in St. Helena—viz., goats. These were introduced in 1513, and multiplied so rapidly, that in 1588 Captain Cavendish states that they existed in thousands, single flocks being almost a mile long. In the year 1709, trees still abounded, though in greatly reduced number, and the governor regarded the goats as of more value than the remaining lumber and firewood! Just a century later another governor reported the final destruc- tion of the forests by the goats which, according to Hooker, ‘‘greedily devour the young plants, and kill the old by browsing on their leaves and bark."’ St. Helena had no native land mammals, and the widely accepted tradition that manatees formerly came to its coast, and that they were captured up to the year 1810, or thereabouts, is erroneous. Mortenson (1933, 417) and Hatt (1934, 560), who have reviewed the evidence, conclude that the alleged manatee was some form of pinneped, such as a fur seal. Fraser's (1934, 33) text and drawing from the manuscript journal of Peter Mundy (circa 1655) suggests that the creature may have been the sea-leopard (Hydrurga leptonyx), although Fraser himself identifies it with the sea-elephant. The bulk of the island’s peculiar fauna is made up of land shells and beetles. Many of the latter are wood-boring weevils, an excellent correlation with the former woodland cover. The fact that there is but a single species of native grass may also be considered an indication that forest once covered the terrain on which many introduced grasses now thrive so well. Land birds of African, Oriental, Australian, European, and American origin have been carried to St. Helena by calling ships, and an extraordinary assort- ment is now acclimated. There is, however, but one endemic form, the ‘‘Wire Bird’’ or plover (Charadrius sanctae-helenae) which the latest reviewer (Peters, 1934, 252) regards as specifically distinct from its nearest relatives in Africa. Together with the almost incredible change in the original cover and ap- pearance of St. Helena, the native marine birds have suffered a vast reduction, both in number and variety. Part of the process has taken place under a vague degree of human observation; the man-o'-war bird, for instance, has disappeared within recent historic time. Much more we can infer only from plentiful remains in the form of bird bones, which have never yet been critically studied. Speaking upon the basis of slight objective evidence combined with zodgeo- graphic likelihood, it is safe to say that in its pristine glory St. Helena was occupied by hordes of shearwaters and other Procellariiformes; by frigate-birds and Red-footed Boobies which nested not only close to the brinks of precipices but also throughout the woodlands, along with the Fairy Terns which still seek sites remote from the sea. Both of the Atlantic tropic-birds doubtless made the island their home, the larger species (Phaéthon aethereus) filling the niches of the sea cliffs with crowded colonies of the sort now to be found in few parts of the world, notably at Razo Island of the Cape Verdes. What other 180 OCEANIC BIRDS OF SOUTH AMERICA birds made up a part of this teeming avifauna we shall perhaps never know in full; the tradition that albatrosses once made.headquarters of the heights of this sub-tropical island may be discounted. Seale (1834, 10) states that in an extensive amphitheatre of St. Helena, be- tween Holdfast Tom and Prosperous Bay Hill, the precipitous sides of which still furnish nesting retreats for sea fow], there are innumerable ancient bird skeletons buried at various depths in the soil. These bones he ascribes to the Red-billed Tropic-bird and the “‘Wandering Albatross."’ The “‘ossuary, in which these fragments are found, extends about a mile in length from the water's edge, . . . and is from ten to ninety feet deep. Remains of a similar description are found near Sugarloaf Hill.”’ According to Blofeld (1852, 195), bird bones are associated with fossil mollusks in various lofty parts of the island. He obtained specimens near Long- wood, at an elevation of 520 meters, on a hillside worn into numerous ravines by heavy rains. The surface, to a depth of two meters, consisted of dark mould. Beneath this was a stratum of grayish brown, friable earth, in which were the bird bones and shells. Owen identified these bones merely as pertaining to ‘marine birds,’’ but others from Turk’s Cap Bay he ascribed to Pleistocene age and stated that they were of the ‘‘petrel kind,’’ some representing the genus Puffinus. Lambrecht (1933, 732) lists the latter under Puffinus lherminieri, a step for which he does not give the evidence. The presence of this cosmopolitan, pan-tropical shearwater at St. Helena would, to be sure, be not surprising, and would help to explain its present-day distribution in the Atlantic. Meliss (1875, 90) likewise refers to remains of extinct birds at St. Helena, and adds, “‘It is difficult also to account fully for the almost total disappearance of some species which once were abundant, such as the Frigate-bird.”’ Many sea birds characteristic of the Antarctic Zone sometimes follow the winds and the cool surface drift to the vicinity of St. Helena. Examples of the Antarctic Fulmar (Priocella antarctica), for example, have been caught on fish- hooks off Speery Island during the month of October, and additional records of species captured or sighted in the vicinity are referred to in the biographies that succeed this section. The resident species, omitting such as the Puffinus which are known only as fossils, are the following, the list being based in part upon collections obtained during the recent expedition of the Cleveland Museum of Natural History. Oceanodroma castro. This petrel nests, according to Meliss, in November. Specimens were collected during the visit of the schooner ‘Blossom’ during October, 1925. Phaéthon aethereus. Once very abundant along the windward coasts of St. Helena, but now greatly reduced in numbers, partly because of former plume- hunting and still more because of the ravages of rats and feral cats. Sula sp.2 There seems to be no exact record of the booby which, according to Meliss (1875, 97), frequents the neighborhood of the island, although seldom coming to land. Fregata sp.2 The man-o’-war bird has long since abandoned St. Helena, and ATLANTIC SUB-TROPICAL ISLANDS 181 it is possible that the specific identity of the form once common there can never be determined. I have thus far failed to locate any record of a St. Helena specimen. The natural inference would be that Fregata minor, which inhabits the sub- tropical island of Trindade, was also the form of St. Helena. However, Ascen- sion is very much nearer than Trindade, lying only 1152 kilometers northwest- ward from St. Helena, and it is at least within the bounds of possibility that Fregata aquila once occupied both islands. Meliss wrote in 1875: “Although there is a part of the southwest coast designated Man-of-War Roost, deriving its name from this bird, and there is still living evidence of its having once frequented even the landing-steps at Jamestown, it is seldom now to be met with, nor is it easy to assign a reason for its disappearance from the Island."’ Sterna fuscata. A migrant species, locally known as the egg-bird, which is said by Meliss to arrive at St. Helena toward the end of the calendar year, and to nest after New Year on some of the outlying islets, such as George's, Speery, and Egg. The Cleveland Museum party, under Simmons, however, collected examples at George’s Island and in the harbor of Jamestown, during September and October. Anois stolidus. The commonest sea bird at St. Helena, often seen resting on the shipping in the roadstead, and nesting in large colonies on Egg Islet and elsewhere. The ‘Blossom’ party collected specimens during October. Gygis alba. Next to the Noddy the “White Bird’’ is the most numerous species at St. Helena. It nests largely in the higher and more inaccessible parts of the main island, occupying ledges of rocky cliffs, and the weathered columnar dikes such as the structures known as Lot’s Wife, etc. Specimens were collected by the ‘Blossom’ party during October. 6. SouTHERN Coasts, GuLF OF LA PLATA TO THE STRAIT OF MAGELLAN In the broad funnel of the Gulf of la Plata, we find a meeting-place between the life of rivers proceeding from the tropics and an oceanic life that suggests the cool south. The great stream pours forth the mud of hot lands, and carries on its brown surface innumerable blue-spiked rafts of “‘camelotes,’’ the South American, bladder-stalked relative of the pickerel-weed (Pontederia). During March, April, and May, the flood season of the Parana, large trees are scattered among the flotsam of lesser plants. It was in May that the ‘Meteor’ encountered remains of such vegetation, as well as yellow water, ‘‘hundreds of miles’’ off the mouth of la Plata (Spiess, 1928, 82). On the other hand, the sea birds that congregate offshore in the wide fringe of mixing water are anything but tropical. Here we find in abundance all of the pan-antarctic petrels, albatrosses, and penguins which have already been listed as sporadic visitors to the coast of Brazil. On the night of July 21, 1832, when the ‘Beagle’ was well to eastward of the estuary, in dirty, squally weather, she was so surrounded by clamorous penguins and seals that the watch was startled into believing the vessel to be driving toward a shore lined with lowing cattle (Darwin, 1933, 80). 60 A’S rou Beck’s te “S.Sebastiao Murphy’s route os Alcatrazes R Colorado R Negro 6. de S. Matias C. Dos Bahias 6. de S. Jorge Comodoro Rivadavia C. Tres Puntas R Deseado 8 Desvelos, Fic. 26. Southern Atlantic coast of South America, with courses of Beck's cruises and local field work during the Brewster-Sanford Expedition, and of the author's voyage in the brig ‘Daisy." [182] SOUTHERN ATLANTIC COAST 183 Both sea and atmosphere in this region are peculiarly subject to strange and violent, and often spectacular, phenomena. The following description of the coastal water just to southward of the gulf is from Darwin's unrevised notebook entry for October 24, 1832, as recently published by Mrs. Barlow: The night was pitch dark, with a fresh breeze. The sea from its extreme luminousness pre- sented a wonderful & most beautiful appearance; every part of the water which by day is seen as foam, glowed with a pale light. The vessel drove before her bows two billows of liquid phos- phorus, & in her wake was a milky train. As far as the eye reached the crest of every wave was bright; & from the reflected light, the sky just above the horizon was not so utterly dark as the rest of the Heavens. And again, at the entrance of the river on April 24, 1833, when much light- ning and an angry sky continued throughout the night, and the royal masthead of the ‘Beagle’ glowed with St. Elmo's fire, Darwin made note of the curious fact that the Rio de la Plata seems to form a nucleus for thunderstorms, which are comparatively rare both to northward and to southward along this coast (1933, 144). Storms from the northward, locally known as ‘‘zondas,’’ are usually de- prived of their rain before reaching this latitude. They are accompanied by a rise in temperature as sharp and remarkable as the lowering which characterizes the storms from the opposite direction, 4. e. the pamperos. The latter blow from the southwest, gathering dust on the pampas, and reach the gulf as a hurricane accompanied by lightning, thunder, and rain. A pampero is generally fore- shadowed by a suffocating day and by heavy clouds to southward and west- ward. The temperature often drops 10° C. within a few minutes after the storm strikes, and the blasts of dust are followed by downpours of rain. The tempest sometimes lasts as long as three days (White, 1881, 1, 15). Hudson (1870, 798) has thus described the way in which duststorms or in- cursions of insects bring birds of the pampas to the edge of the gulf, the gulls referred to being Larus maculipennis, although in this citation the author used another specific name: Whenever there is a dust storm in the vicinity of Buenos Aires myriads of gulls . . . appear flying before it. This is always the case even when a gull had not previously been seen for months. A dust storm is always preceded by a long drouth, which means that the water-courses along which the gulls would ordinarily live have been dried up. In seasons when grasshoppers abound, flocks of the gulls also appear working as veritable armies and sometimes coming in such multitudes that they free entire districts from the devastating storms of the insects. The line of birds some- times presents a front of several thousand feet, with a breadth of 60 or 80. Along this line of battle the cries of the birds produce a loud incessant noise and all the time the hindmost birds are flying over the flock and alighting in the front ranks so that the whole body is steadily advancing. The entrance of the great bight of the Rio de la Plata is 193 kilometers wide from Punta del Este in Uruguay to Cape San Antonio on the outer coast of the Province of Buenos Aires, Argentina. The gulf receives the water of six rivers, comprising the second largest river system in the world. In ancient times the extent of the inlet was very much greater than at present, as is proved by the deposits of marine shells far up the Rio Parana, but the size of the old arm of 184 OCEANIC BIRDS OF SOUTH AMERICA the sea has decreased proportionately with an uplift of the land. The estuary as a whole is remarkably shallow, and the silt carried by the river water nor- mally discolors the ocean for 125 kilometers or more from the coast. Islands at the mouth of the Parana are formed of this recent sediment and, according to Burmeister (1876, 283), one finds in these deposits remains of contemporary marine animals, from which he infers that the Atlantic penetrated this far even within recent geologic times. At the present day much of the outer body of water still pertains to the ocean, but the inner part, extending from Punta de Espinillo, west of Montevideo, and Punta de las Piedras on the Argentine side, contains mostly fresh water and has lost practically every marine attribute. — The outer gulf therefore represents an expanded bay in front of the triangular estuary of the combined rivers. Except in freshet season, the water is here often clear and has considerable salinity, which becomes rapidly reduced inside Montevideo. When en route from the north, the first land seen by the traveller is usually the Lobos Islands which, in turn, are likely to be presaged by thou- sands of seals playing in the open sea (Morong, 1892, 479). Here the pilot may be taken aboard and shortly afterwards the Banda Oriental, or ‘‘Purple Land,” comes into view. The shores nearly everywhere are bluffs of granite and other metamorphic rocks, with the rolling grasslands beyond. Cunningham (1871, 44) describes his first glimpses of this part of the coast as follows: On the 23d of Nov. land was reported soon after sunrise. As we steamed up the estuary of the Plate, the low land on the northern shore, between Maldonado and Monte Video, brought to my remembrance my first sight of the Dutch coast, to which it bears a considerable resemblance in respect of its sand-dunes, with the low-lying, mostly treeless country beyond them, dotted here and there with windmills. A large wasp flew on board in an exhausted condition, and about the same time the rigging of the ship became covered with immense quantities of cobweb. Every- where long delicate threads could be seen streaming out into the air, and a considerable number ot their constructor, a minute reddish-brown spider, were to be observed associated with them. This curious phenomenon is of frequent occurrence in the Plate on a fine day after stormy weather. In the same place a century ago, the rigging of the ‘Beagle’ likewise became coated and fringed with gossamer webs of the aéronaut spiders, some of which, Darwin calculated, must have come through the air a hundred kilometers. Many butterflies, also, are often seen here, and dead insects sometimes pepper the surface of the water, while it is not unusual for many kinds of land birds to rest On passing ships. Along the rocky northern shore of the gulf, as far westward as Montevideo, there are large areas inaccessible to modern steamers. Even in the deeper ex- panses toward mid-channel numerous banks, covered by only a few meters of opaque water, have been the cause of countless wrecks. The beginning of the inner, triangular part of the gulf can at once be recognized by the muddy hue of the water and its lessened salinity, although the prospect is still that of a sea, with no shore yet visible to southward. A stiff breeze from the latter direc- tion will quickly pile up sluggish waves, which Darwin likened to moving mountains of mud. He also made the following acute observation regarding the interrelationships of the continental waters and those of the ocean (1933, 82): SOUTHERN ATLANTIC COAST 185 The fresher, discolored water from its less specific gravity floats on the surface of the salt; this was curiously shown by the wake of the vessel, where a line of blue might be seen mingling in little eddies with the adjoining fluid; in this case instead of stirring up the mud, it was the reverse and stirred up the clear water. The ridge of sand marking the Banda Oriental terminates in cliffs opposite the island of Martin Garcia. From this point, if the traveller has clung to the northern shore, the Argentine coast may appear magically, under a mirage, with the trees suspended in air. The southern side of the gulf presents a marked contrast with that on the north. It is rockless and without hills, almost level, and so low that it does not become visible from a deck until it is very closely approached. Its general elevation is from 2 to 7 meters above the sea, and there is but one kind of native tree, known as the ‘‘ombu,”’ which sometimes grows to a height of 12 meters and is commonly buttressed around the base with large knees. From the vicinity of Magdalena, all the way out to Cape San Antonio, there are few landmarks on the low shores other than scattered groups of these trees. Along the semi- circular beach of the Ensenada de Samborombon, between Punta de las Piedras and Cape San Antonio, are very extensive flats and shallows, with water less than 10 meters in depth at distances of 6 to 7 kilometers from shore. The Rio Salado, largest stream of the pampas, reaches the sea in the middle of this stretch, but is entirely inaccessible from the gulf; as a rule the depth of water is not more than a meter at its mouth. The whole water front here, as also the grass- lands stretching toward Lake Chascomis and numberless other bodies of fresh water, is a famous resort of ducks, swans, flamingoes, the Coscoroba, and many species of shore birds, including native forms as well as migrants from North America and the Arctic. Gibson (1918, 363) states that the coastal land is here so young that remains of six or more whale skeletons may still be seen on one hacienda. Floods sometimes cover much of it, as they did during October, 1913. The country just inside, or to westward, of Cape San Antonio is thus described by Gibson who, at intervals of forty years, has written two delightful papers about the birds of the region: . .. . here are the “‘rincénes,’’ a maze of islands, peninsulas, and tidal creeks, with count- less ramifications. The surface of the mainland is here recent and quaky, with a softer subsoil. The innumerable creeks are called ‘‘cangrejales’’ from the myriads of small burrowing crabs inhabiting their banks. Inland the cangrejales merge with freshwater fens. There are small areas of natural woods on rises of ground in this district, although the average altitude is probably not more than six feet above sea level. At night the uncanny shrieks of the viuda loca (limpkin) may be heard from the swamps. Outside Cape San Antonio the beaches are bathed by blue, saline water, that is, ‘‘blue’’ compared with the muddy water of the rivers. As a matter of fact, the ocean surface over the entire coastal shelf between Uruguay and Cape Horn, and offshore to a point well east of the Falkland Islands, is prevailingly greenish, and quite unlike the ultramarine sea along the central coast of Brazil. Southward from Cape San Antonio the sand hillocks gradually give way to a dead-level line of cliffs about 10 meters high, an indication of the impressive 186 OCEANIC BIRDS OF SOUTH AMERICA and considerable recent elevation of the coast, which increases southward into Patagonia. Behind the bluffs stretches the level and unbroken pampa except where, at long distances apart, a line of bare hills becomes visible from sea. Such hills, an occasional bend or corner, an islet or promontory, or the narrow gorge of a river cut through the table-land, henceforth make the landmarks of an otherwise almost featureless coast. In the summarized narrative of the Brewster-Sanford Expedition, near the beginning of this book, I have referred to Mr. Beck's profitable visit at Mar del Plata in the austral spring of 1914. Making headquarters at this popular water- ing-place on the outer coast of the Province of Buenos Aires, just north of Cape Corrientes where the shore line turns westward toward El] Rincon, he divided his time between land and sea, according to each day's weather, and had full opportunity to observe the wealth of local birds. Between September 18 and October 27 he collected about 375 specimens, comprising 25 to 30 dry- land forms and a larger number of ocean fowl, marsh birds, and shore birds. Much of the country was flooded during Beck’s stay; his notes for late September tell of cornfields ankle-deep or knee-deep in water, with thousands of ducks feeding close to plowmen working under difficulties, and almost oblivious even of gunners. The bulk of the ducks were Brown Pintails, but there were also plenty of Widgeon, Shovellers, Teal, and others. A good proportion of the water birds and waders were beginning to nest, as were also two kinds of tinamou, Short-eared and Burrowing Owls, and many lesser plains and thicket birds. Mingling with and feeding with these resi- dents, however, were numerous species at the opposite physiological pole of their annual cycle, namely the migrant Charadriiformes from North America. This peculiar combination, to be best observed, perhaps, in this very part of South America, is a phenomenon unknown anywhere in the holarctic continents. Mr. Beck collected a goodly representation of migrant snipe, sandpipers, plovers, etc., only one example of which needs special mention. On September 29 a shore bird which he took to be an Eskimo Curlew (Numenius borealis) flew past him. Beck whistled an imitation of the Hudsonian Curlew’s call, which had no effect. He then gave the call of the Black-bellied Plover, whereupon the Curlew at once turned and flew into range. It proved to be an Eskimo Curlew, with its stomach characteristically full of insects. From several even more recent records, we know that this “‘virtually extinct’’ arctic species probably still visits the pampas. The general composition of the bird life back from the seashore on the lake- dotted plains of the Province of Buenos Aires is, in many respects, suggestive of the avifauna of somewhat similar prairies in the interior of North America. The analogy is interesting in view of the fact, stressed elsewhere, that the Argentine pampas correspond geographically with the central part of the north- ern United States, rather than with the eastern seaboard. In the following partial list of pond and swamp birds of Buenos Aires, mostly collected by Beck or mentioned in his journal but added to from the excellent monograph by Wetmore (1926), such resemblances will be evident. SOUTHERN ATLANTIC COAST Native SPEcIEs Chilean Grebe. Colymbus chilensis. Pied-billed Grebe. Podilymbus podiceps antarcticus. Night Heron. Nycticorax nycticorax hoactli. Azara’s Least Bittern. Ixobrychus involucris. Maguari Stork. Euxenura galeata. Chilean Flamingo. Phoenicopterus chilensis. Screamer. Chauna torquata. Black-necked Swan. Cygnus melancoriphus. Fulvous Tree Duck. Dendrocygna bicolor bicolor. White-faced Tree Duck. Dendrocygna viduata. Coscoroba. Coscoroba coscoroba. Cinnamon Teal. Anas cyanoptera. Gray Teal. Anas versicolor. Yellow-billed Teal. Anas flavirostris. South American Pintail. Anas spinicauda. Southern Bahama-duck. Anas bahamensis rubrirostris. Chiloé Widgeon. Mareca sibilatrix. Red Shoveller. Spatula platalea. Black-headed Duck. Heteronettd atricapilla. Rosy-billed Duck. Metopiana peposaca. Southern Ruddy Duck. Oxyura vittata. Limpkin. Aramus scolopaceus carau. Magellanic Rail. Rallus limicola antarcticus. Argentine Sora. Porzana spiloptera. Great-footed Coot. Fulica armillata. White-winged Coot. Fulica leucoptera. Red-fronted Coot. Fulica rufifrons. Painted Sandpiper. Nycticryphes semi-collaris. Teru-teru. Belonopterus chilensis lampronotus. Collared Plover. Charadrius collaris. Paraguayan Jack-snipe. Capella paraguaiae paraguaiae. Black-tailed Stilt. Himantopus himantopus melanurus. Kelp Gull. Larus dominicanus. Gray-hooded Gull. Larus cirrocephalus. Brown-hooded Gull. Larus maculipennis. Trudeau's Tern. Sterna trudeaui. NorTHERN-HEMISPHERE MIGRANTS Golden Plover. Pluvialis dominica dominica. Eskimo Curlew. Numenius borealis. Hudsonian Godwit. Limosa haemastica. Lesser Yellowlegs. Tringa flavipes. Greater Yellowlegs. Tringa melanoleuca. 187 188 OCEANIC BIRDS OF SOUTH AMERICA Spotted Sandpiper. Actitis macularia. Robin Snipe. Calidris canutus rufus. Sanderling. Crocethia alba. White-rumped Sandpiper. Erolia fuscicollis. Baird's Sandpiper. Erolia bairdtt. Pectoral Sandpiper. Erolia melanotos. Buff-breasted Sandpiper. Tryngites subruficollis. Wilson's Phalarope. Steganopus tricolor. During the southern-winter season there is also a migration into the Province of Buenos Aires from more southerly latitudes and from the highlands to west- ward. In this movement such shore birds as the Falkland Plover (Charadrius falklandicus) and the Seed Snipe (Thinocorus rumicivorus) take patt. In certain sections of the Province of Buenos Aires, and still more toward the borders of Patagonia and beyond, some of the standing bodies of water are salt lagoons. These, too, are well peopled with birds. Doering (1881, 19) lists as characteristic species the Bahaman, Yellow-billed, and Cinnamon Teals, the Shoveller, Chilean Flamingo, Falkland Plover, Stilt, and Kelp Gull. The last is typically a coastal bird, but Wetmore (1926, 133) found it on the pampas nearly 200 kilometers from the ocean, which is the less surprising since it is said to nest about some of the larger Andean lakes in Argentina. Flamingoes noted by Wetmore in the salt lagoons seemed to be feeding upon a species of brine shrimp (Artemia). The foregoing lists, which might readily be extended, refer altogether to birds found in various types of terrain behind the shore. But Mr. Beck also carried on much field work from boats during October, 1914, making several trips to fishing-banks about 30 kilometers off the coast, where the craft of the Mar del Plata fleet were crowded together so that they almost touched one another. It was on these banks, usually when fish were being hauled in fast, that the greatest number of pelagic birds was seen. Here, as on the land, there proved to be an extraordinary intermingling of species from opposite ends of the Atlantic Ocean, but in this case even the southern-hemisphere forms were mostly at a considerable distance from their nesting grounds. The species collected, or in one or two instances merely observed, by Beck between the roadstead of Mar del Plata and his farthest point from shore were the following: SoutH ATLANTIC SPECIES Magellanic Penguin. Spheniscus magellanicus. Royal Albatross. Diomedea epomophora. Black-browed Albatross. Diomedea melanophris. Giant Fulmar. Macronectes giganteus. Cape Pigeon. Daption capensis. White-chinned Petrel. Procellaria aequinoctialis. Greater Shearwater. Puffinus gravis. Sooty Shearwater. Puffinus griseus. SOUTHERN ATLANTIC: COAST 189 Wilson’s Petrel. Oceanites oceanicus. Chilean Skua. Catharacta skua chilensis. Brown Skua. Catharacta skua, subspecies? Kelp Gull. Larus dominicanus. Brown-hooded Gull. Larus maculipennis. South American Tern. Sterna hirundinacea. MIGRANTS FROM THE NorTH ATLANTIC Manx Shearwater. Puffinus puffinus. Parasitic Jaeger. Stercorarius parasiticus. Long-tailed Jaeger. Stercorarius longicaudus. Royal Tern. Thalasseus maximus. Common Tern. Sterna hirundo. On several days, according to Beck's journal, the two species of jaegers from breeding grounds in the arctic tundra were the commonest birds along the coast of eastern Buenos Aires. Sooty Shearwaters from the Cape Horn region, Greater Shearwaters from Tristan da Cunha, and Manx Shearwaters from Europe or Bermuda, perhaps, sometimes formed indiscriminate bands. In general, Beck thought, the flocks of migrant shearwaters that seemed to be bound southward kept much farther from the Argentine shores than similar streaming flocks do from the shores of California and Chile. From Mar del Plata and Cape Corrientes the low coast, which is generally swept by a long-shore surf driven before winds off the pampas, turns westward toward El Rincon, the angular concavity in which Bahia Blanca lies. The Sierra de la Ventana, on the northern side, rises to an altitude of 1250 meters, the greatest elevation on the Atlantic coast between Brazil and Tierra del Fuego. The entrance to Bahia Blanca and the whole western shore of El] Rincén are lined by low, flat, marshy islands, intersected by creeks. In the southern part of the bight, sandbars extend up to 25 kilometers to seaward, and even the great rivers from the Andes, like the Colorado and the Negro, are partially blocked at their mouths with banks which are pounded by tremendous surf when the wind swings to an easterly quarter. Darwin (1839, 93) thus describes the sheltered inner harbor of Bahia Blanca: The wide expanse of water is choked up by numerous great mud-banks, which the inhabitants call Cangrejales, or crabberies, from the number of small crabs. The mud is so soft, that it is im- possible to walk over them, even for the shortest distance. Many of the banks have their surfaces covered with long rushes, the tops of which alone are visible at high water. On one occasion, when ina boat, we were so entangled by these shallows, that we could hardly find our way. Noth- ing was visible, but the flat beds of mud: the day was not very clear, and there was much refraction, or as the sailors expressed it, “‘things loomed high."’ The only object within our view, which was not level, was the horizon; rushes looked like bushes unsupported in the air, and water like mud-banks, and mud-banks like water. The islands along the western shore of El Rincén, such as Bermejo, Trinidad, and Verde, as well as others south of the double mouths of the Rio Colorado, 190 OCEANIC BIRDS OF SOUTH AMERICA are all low, occasionally more or less overflowed, and suitable only for marsh birds. Little is specifically known of their fauna, but the southernmost islet in San Blas Bay, just north of Punta Rubia, bears the name of Isla de la Gama or ‘Deer Island.”’ The Rio Negro marks the northern boundary of Patagonia, a word which now has only traditional political meaning, for it finds no place among the modern geographic divisions of Argentina. Nevertheless, it will always main- tain a regional and zo6geographic significance, for Patagonia is the home of many organisms not found outside its borders. For example, while the Common Rhea (Rhea americana albescens) ranges southward only to the Rio Negro, it is replaced between the south bank of that river and the Strait of Magellan by a bird of a different genus, Darwin’s Rhea (Pterocnemia pennata pennata). From another distributional point of view, we may cite the Condor (Vultur gryphus), a mountain bird which regularly descends to sea level in desert districts. This species crosses the continent from the Andes of northern Argentina, and reaches the Atlantic coast only southward of the Rio Negro from where, like Darwin's Rhea, it ranges to the Strait of Magellan. Related instances might be multiplied. The Rio Negro is shallow and treacherous at its mouth, like all other rivers of this coast. It enters the sea just north of the second of the three deep bights between the Rio de la Plata and the Strait of Magellan, namely the Golfo de San Matias, This inlet marks the beginning of the Patagonian table-land at the coast, for the whole western border of the gulf is high and bold. To north- ward and eastward are plains which slope almost imperceptibly toward the Atlantic, but at San Matias it is as though the ocean had cut into the North American continent as far as the meridian of Chicago, to which, in such a fanciful geographic comparison, the excellent port of San Antonio might be likened. In the Gulf of San Matias we also reach the northern limit of the ancestral breeding range of the Magellanic Penguin on the east coast of South America, although the vast majority of the birds are now confined to more southerly nesting localities. The present or former existence of seals and sea-lions in abundance is indicated by such names as Banco Lobos and Punta Lobos, which appear, not once but several times, on charts of the adjacent shores. We have now reached, by the way, one of the world’s ‘guano coasts,’’ though of far less importance than that on the opposite side of the continent. Guano, derived from the excreta of penguins and other birds, and doubtless also the inferior product of sea-lions, was formerly “‘mined’’ at many Patagonian localities between Bahia Blanca and Rio Gallegos. The southern side of the Gulf of San Matias is formed by the Peninsula de Valdez, which is nearly constricted off from the coast of Chubut by the smaller gulfs of San José and Nuevo. The expanded termination cf the peninsula, which is 85 kilometers in length from north to south, differs little from the mainland, being flat country with shallow depressions, covered with the usual stunted vegetation and inhabited by the usual birds and mammals of the pampas. The Golfo Nuevo, on the southward side, was formerly supposed to be stagnant SOUTHERN ATLANTIC COAST 191 in its depths, and hence of peculiar physico-chemical character throughout. The ‘Meteor’ party recently discovered, however, a deep and hitherto unknown channel through the barrier, indicating that new supplies of Atlantic water constantly reach the gulf at all levels (Spiess, 1928, 169). Along the Patagonian coast the flood-tide comes in a wave from the south, and on the outer front of Valdez Peninsula it advances with great violence, causing overfalls which are dangerous to small craft. The onrushing water pours with a speed attaining 11 kilometers per hour into the narrow inlet known as Valdez Creek, on the eastern side of the peninsula, and into the Gulf of San Matias and connecting bodies of water. The average rise and fall is 6 meters in this region. Southward it becomes less, being about 3.6 meters at Puerto Des- eado, after which it increases again toward the Horn, attaining a range of 7.3 meters at Santa Cruz and still more near the Strait of Magellan. The entire Atlantic coast between the Province of Buenos Aires and the Strait lies within the zone of prevailing westerly winds. From September to June these are interrupted at frequent intervals by southeasterly gales which are the winds that bring the radns and raise the heaviest seas. Between iatitudes 40° and 50° S., that is from El Rincén to Puerto Santa Cruz, the seaboard is characterized by notable uniformity of climatic conditions. In many parts of this stretch of coast it is injurious for wooden ships to lie at anchor for even brief periods during the summer season, so desiccating is the power of the rarely clouded sun in the practically complete absence of either rain or dew. Darwin (1933, 220) was extraordinarily impressed by the strong westerly breezes, with dry air and a brilliant sky, encountered at the mouth of the Rio Santa Cruz. He never ceased to regard it as marvelous, moreover, that at the distance of less than 200 kilometers to southward, such astonishing weather conditions were replaced by almost constant cloudiness, rain, hail, sleet, and snow. The essential geographic trait of Argentina is, indeed, the great transfor- mation in the climatic régime from north to south. In the northern and central parts of the republic, the rainfall diminishes regularly from east to west. To- ward the south, on the contrary, as soon as one has entered the zone of perpetual west winds, the rains are borne from the west and give out toward the ocean. In these latitudes, as one leaves the slopes of the Andes and goes toward the Atlantic, aridity steadily increases, forests disappear, the steppe itself gives out; the rivers, carrying only the water of their remote montane sources, receive no affluents from the pampas, which have either an interior-basin drainage or are without surface-flow of any sort. Toward the eastern ocean the country bears more and more the stamp of the desert. San Antonio, on the Gulf of San Matias, receives but 180 millimeters of rainfall annually, and a similar supply holds for a long distance southward. Beyond the Rio Santa Cruz, however, precipitation begins to increase again, this being the latitude in which the cordillera drops somewhat, making a gap through which moisture-bearing winds are free to blow across the continent from the Pacific. Puerto Gallegos receives 400 milli- meters of rainfall; Ushuaia 500. A system of winter rains is general on the plateau, but less regular at the coast. The winter maximum is here interrupted 192 OCEANIC BIRDS OF SOUTH AMERICA by a short dry period (August at San Antonio, July and August at Bahia Cama- rones, June at Puertos Deseado and Santa Cruz), which is marked by a truce of the west winds. The coast is washed by cold ocean waters and the land breezes cause upwelling from the depths, resulting in mists and fogs, which do not penetrate far inland and which recall the Peruvian gartas. The blankets of fog, while often very extensive over the offshore waters, are likely to be exceedingly shallow, so that the masts and upper spars of ships project through them, a , phenomenon mentioned in many accounts of voyages. The fogs are only one instance of the erratic possibilities of Patagonian coastal climate, despite its prevailing uniformity. Broiling sunshine and light breezes may be suddenly succeeded at almost any season by cold blasts that bring masses of low clouds and bursts of rain. Periods of drought, such as are characteristic of the whole shore between Bahia Blanca and the Strait of Magellan, are likely to be abruptly broken by squalls and fogs. As the humidity increases to southward of Santa Cruz, the rainy season also changes. At Rio Gallegos the wettest month is December; at Ushuaia the steady rains last from September to March. The volume of river water is entirely determined by distant snowfall. From the Colorado and the Negro southward to the Santa Cruz and the Gallegos, all the streams have their freshets at the end of springtime; in autumn those which do not come from the Andes dry up long before they reach the ocean. As has perhaps been sufficiently inferred, the Patagonian climate in the region between the Rio Negro and the wet zone of the southern tip of the continent 1s not only arid but is also rude. Violent westerlies rage clear to the Atlantic, and on across it. Simpson (1934, 184) humorously remarks that it is important ‘to keep papers from blowing away, out of consideration for the people in Africa!’’ Along the coast the mean temperature lowers southward a little less than one degree Centigrade for each degree of latitude. Thus it is 14.5° at San Antonio near 41° S.; 8.5° at Santa Cruz, in 50° S.; 4.4° at Ushuaia, close to 55° S. The average temperature of the summer months is more sharply reduced, being 21.4° at San Antonio; 14° at Santa Cruz; and 9.2° at Ushuaia. Grain does not ripen south of Chubur. Along the Patagonian coast from the neighborhood of Valdez Peninsula southward, we begin to encounter masses and marine hedges of the giant kelp (Macrocystis pyrifera). This is a seaweed of great importance in the ecology of most forms of sub-antarctic neritic life. It is also a living index of the general range of ocean temperatures and of the distribution of many organisms, such as penguins and the southern pinnepeds, which select cool and richly nutrient waters. The first extensive fields of kelp are found, perhaps, in Desvelos Bay, a short distance south of Puerto Deseado. From here southward it grows abundantly along all sub-antarctic shores, and about the northerly antarctic islands, such as South Georgia. Uprooted strands of kelp may, of course, be cast by the sea on beaches far removed from the point of origin. Admiral Spiess mentions finding pieces 17 meters in length along the African coast, supposedly carried from Tierra del Fuego by the west-wind drift. The west coast of South SOUTHERN ATLANTIC COAST ’ / 493 America, because of conditions produced by the Humboldt Current, is favorable for the extension of kelp far into the tropics; so that along the continental border as a whole it is, in effect, one of the symbols of the pan-antarctic littoral, as discussed heretofore. Cunningham (1871, 60) gives the following picture of the giant kelp and of its place in nature: This wonderful plant, the most gigantic Alga known, exists in vast beds around the coasts of Patagonia, Tierra del Fuego, and the Falkland Islands, in general growing in depths of from six to twenty fathoms, and is of the greatest service to the navigator as an indication of the presence of rocks to be avoided by him. From a branching root, in the intricacies of which small Molluscs, Crustacea, Echinoderms, and Annelids nestle, arise small fructiferous bladderless submerged fronds, and long slender stems, which reach the surface of the water, and there give off hundreds of elon- gated elegantly-shaped jagged-edged fronds, varying in length from four to six inches to one or two feet, each provided with a pyriform air-vesicle at the base. These fronds, derived from one another by a process of vertical splitting, spread out on the surface of the water like so many banners, the manner in which they are directed being an infallible index of the ebb and flow of the tide, and I know few more beautiful sights to be witnessed, than by leaning over the gunwale of a boat on a calm day, and gazing through the clear depths of these submarine forests, in which fish swim about as birds fly through the trees of a wood. Vallentin (1924, 378) writes that at the Falklands . where, in great patches, this weed luxuriates, it forms such a dense mass that I have often experienced great difficulty in forcing my boat through it during low water. Frequently, when out collecting in my dinghy, I have comfortably ridden out a heavy squall by making fast the bows to a substantial growth of seaweed. Most of the roots of Macrocystis are attached to rocks or stones of varying sizes, but in many of the protected fiords they seem to grow freely on the muddy bottom. In the interstices of these tangled roots hosts of forms lie hidden, and one or more specimens of the fish Lycodes can nearly always be found therein. A list of the inhabitants of one of these roots would include the names of almost all the invertebrates found round these islands; while on the fronds abound hydroids and bryozoa of various kinds. Crawshay (1907, xxii) states that during the expedition of the ‘Erebus’ and ‘Terror’ kelp strands more than 200 meters in length were found. He also quotes Sir Joseph Hooker, the botanist of the expedition, as follows: In the Falkland Islands, Cape Horn, and Kerguelen’s Land, where all the harbours are belted with its masses . . . , it generally rises from eight to twelve fathom water, and the fronds extend upwards of one hundred feet upon the surface. We seldom, however, had opportunities of measuring the largest specimens, though washed up entire on the shore; for on the outer coasts of the Falkland Islands, where the beach is lined for miles with entangled cables of Macrocystis, much thicker than the human body, and twined of innumerable strands of stems coiled together by the rolling action of the surf, no one succeeded in unravelling from the mass any one piece upwards of seventy or eighty feet long; as well might we attempt to ascertain the length of hemp fibre by unlaying a cable. Southward from the Valdez Peninsula the coast continues high in front of country which has long been farmed by Welsh settlers, and which is at least more populous than most of Patagonia. The coast, and the valley of the Chubar, are notable in ornithological annals because of the collecting and writings of Durnford. By way of picturing the abundance of resident sea birds, I may quote this author's description of a colony of terns (Sterna hirundinacea) made at Tombo Point, the southern headland of Janssen Bay, just beyond latitude 44° S. 194 OCEANIC BIRDS OF SOUTH AMERICA I was prepared when I visited this place in December to see a considerable quantity of birds; but the number that met my eyes fairly staggered me. These nests cover an area about 150 yards square. Allowing three nests and five eggs for every square yard (a very moderate computation, it being difficult to walk without treading on the eggs), we arrive at the extraordinary number of 67,500 nests, 135,000 birds, and 112,500 eggs; and, wonderful as these figures may appear, I feel sure that I eee rather understated than overstated the numbers (Durnford, 1878, 404). The Rio Chubit, which enters the sea in the middle of this stretch of coast, is narrow and shallow at its mouth, and approach to the hinterland is made mostly by way of Puerto Madryn, in the Golfo Nuevo. In fact, the Rio Santa Cruz, far to southward, is the only Patagonian stream accessible to large vessels, and this more because of the great tidal rise in its estuary than of mean depth in the channel. All of this coastland has been progressively uplifted between middle Tertiary time and the present day. Several distinct systems of coastal terraces along the Atlantic have resulted from the emergence, which at the edge of the sea has apparently been greatest in the neighborhood of Puerto Deseado. The more recent terraces contain fossils of species now living in the adjacent ocean, indicating a relatively late lowering of the marine level (Rudolph, 1934, 258). Feruglio (1933, 1) recognizes six successive terraces, as follows: VI. 5-6 meters . Mollusks all of local recent species. V. 812m. . . Mollusks all of local recent species. IV. 15-30m. . . Mollusks all of recent species, but with some austral and Chilean elements suggesting colder waters. Ill. 35-60m. . . Mollusks mostly recent species, but with some north- ern elements suggesting warmer waters. II. 115-140 m. . . Some extinct mollusks. I. 165-185 m. . . Many extinct mollusks. The shore everywhere between the well-separated ports presents a wild and lonely appearance, enlivened chiefly by birds that inhabit the strands and cliffs and the noisy herds of sea-lions on such islets as those of Leones and Toba, south of Bahia Camarones, as well as at favored places on rocks or shingle beaches of the mainland itself. The roads parallel to the sea usually run some distance inland, and practicable paths down the bluffs are far apart, sometimes 15 kilometers or more. The few rivers enter the ocean through deep and steep- cut valleys, and in some places altitudes of 200 to 300 meters rise abruptly from the shore. Simpson (1934, 273) describes a recent trip by motor, southward on roads varying from indifferent to execrable along the coastal terraces of the Golfo de San Jorge, from Comodoro Rivadavia to Puerto Deseado. He refers to a very large ‘‘loberia’’ or seal rookery on a beach of the gulf which, for some incom- prehensible reason, has been persistently singled out by the sea-lions throughout a long period, in preference to identical sections of beach on either side. Here an old local sea-lion fishery has now developed into an elaborate modern whaling station, in which the machinery is operated by electricity, and from where both SOUTHERN ATLANTIC COAST 195 local sealing and offshore whaling are conducted at respectively appropriate seasons. Between the Rio Chubdt and the Rio Deseado, and most notably along the northern half of the Golfo de San Jorge, the coast follows great mesetas, or high pampas, which are elongate, relatively narrow, and parallel with the shore line. The only marked gap is the Gran Bajo Oriental, between Comodoro Rivadavia and Caleta Olivia. The region back of the mesetas (aside from en- closed basins) is drained by the Senguerr-Chico system and the Deseado. Near longitude 69° 30’ W. these are less than 80 kilometers apart, but they then turn northeast and southeast, respectively, and their waters reach the sea (via the Chubiat in the northern system) some 500 kilometers apart in a straight line. The lower courses of each system are parallel with the coast, west of the mesetas, for some distance. Thus the Rio Deseado is little more than 30 kilometers from the sea at a point from which it flows more than 150 kilometers before reaching it. In like manner, the Rio Chico is only 45 kilometers inland in a straight line at a point 275 kilometers from its mouth. This meseta barrier has had a very marked influence in the history of the region, and the country behind it was almost the last to be explored and settled. The usual shore in this section, roughly that of the Golfo de San Jorge, is marked by three features: restinga, beach, and cliff. The restinga is a broad wave-cut platform in the country rock (Eogene tuffs, Patagoniano, etc., here always Miocene or earlier), and its level is between tides but nearer low tide. Its surface is broadly plane, but very rough in detail, with many pools in which fishes and other marine animals survive between tides. The rocky projections are usually covered with mussels. Birds, especially gulls, congregate here in enormous numbers at low tide. Beach débris is not commonly found on the restinga. The beach usually begins abruptly on the shoreward side of the restinga, and extends to the base of the cliff. It may be broad and nearly level, but more commonly is narrow and steep, close to the angle of repose for its material. Modifications, such as ridges and beach cusps, are common. The usual material is shingle, derived from the famous Tehuelche Gravel that covers the pampas of Patagonia, but reduced in size and often brightly polished by wave-action. This grades into fine sand, as at the Balneario, south of Comodoro. At the points and along some of the indentations the sea cliff is high and abrupt, and is cut in the native rock, like the restinga, the structure being that of a cut-notch or step, with débris, the beach, piled against its riser. Along the larger and older indentations, however, such as that of Bahia Solano, the true cut-cliff is low or absent, and the beach abuts against a series of older marine terraces (Simpson, MS). The Golfo de San Jorge terminates at Capes Tres Puntas and Blanco, beyond which there are great shoals along the coast running nearly due south toward Deseado, the Port Desire of the early British explorers and one of the oldest settlements in Patagonia. The river to which the town owes its existence is little more than a trickle, but it opens into an estuary to which vessels can 196 OCEANIC BIRDS OF SOUTH AMERICA scud for shelter, rare enough on this coast. Just south of Puerto Deseado are the famous Penguin Islands, described hereafter in the biography of the Magel- lanic Penguin. The mainland cliffs are also occupied by many nesting water- fowl, and the Isla del Cafiadén del Puerto, in the very mouth of the river, is probably the northernmost breeding station of the Gray Cormorant (Phala- crocorax gaimardi) on the east coast of America (Renard, 1931, 412). Other islets, such as Shag Rock, off Hilly Point, where the coast turns southwestward, and Bird and Flat Islets, which lie in front of the great table-land just north of Port San Julian, are likewise the headquarters of an important aggregation of marine birds which are associated with definite oceanographic conditions in South America, and several representatives of which we are now about to en- counter for the first time during the course of our circumnavigation. All along this water front the table-land rises just above the breaking waves, being interrupted in places by broad arroyos which are dry and salt-encrusted during much of the year. At certain points, indeed, whole blocks of the plateau have weathered away, so that only isolated natural turrets and castles of strati- fied material remain. South of Puerto San Julian, between latitudes 49° and 50° S., steep white cliffs, with an average height of about 100 meters, loom like a wall from the sea, which at high tide washes their bases but at low water leaves a broad front of shingle and mud. Such Patagonian cutbanks are famous in the history of palaeontology, from the days of the ‘Beagle’ to the present, fossils being exposed at nearly all levels. Below the sea cliffs, moreover, the tidal flats make a great natural sorting-pan for the detritus of a world that is no more, as indicated by the mineralized bones and trees, and other manifestations of Tertiary life. The estuary of the Santa Cruz is a relatively sheltered and commodious arm of the ocean, navigable well inland, and with sea-lion and penguin islets out of sight and sound of the open Atlantic. Spring tides here sometimes attain a maximum rise of nearly 18 meters (Riggs, 1926, 537). The gorge or fiord drops abruptly from an extremely level and sterile plain. The river itself is geographi- cally one of the most important, and historically perhaps the most interesting, of all the Patagonian streams. Here, in April, 1834, Fitz Roy and Darwin, with their boat parties, made a traverse across the pampas, and rather tragically turned back just as they were on the verge of discovering Lago Argentino. After one day's journey up the river the ‘Beagle’ found itself practically beyond tidal influence. Between here and the farthest point reached during the boat- trip of several weeks, the stream holds a breadth of from 300 to 400 meters and a depth in the center of the channel of about 5 meters. Most remarkable is the constant rapidity of the current, which runs at a rate of from 7 to 11 kilometers per hour, the water being always of a fine blue color, very slightly milky. The entire valley formed by the age-long erosion of the stream maintains a width of from 8 to 15 kilometers and leads very directly westward, bounded by horizontal plains of from 100 to 150 meters in elevation. In the bottom of this great canyon the river itself follows a meandering course (Darwin, 1933, 221). SOUTHERN ATLANTIC COAST 197 pe “et5>—) Falkland I$ ‘tir a? /¢ anle’ SE Povey Shag Rocks BJ al. de los Estados South Georgia” ae Beck's route 6 di Murphy's route I$ Ildefonso y | I 70 = ‘I8 Diego Ramirez 50 Fic. 27. Southernmost South America and the associated islands. For detailed maps of the latter see figures 37 (Tierra del Fuego), 28 (Falkland Islands), 31 (South Georgia). In addition to the tracks of Beck’s and Murphy’s travels, local American Museum field work is indicated by figures in circles: 4, Chapman, reconnaissance in Tierra del Fuego and western Patagonia; 5, Correia, at South Georgia. The coast between Santa Cruz and the eastern entrance of the Strait of Magel- lan, and beyond toward the tip of Tierra del Fuego, requires no detailed topo- graphic description, because of its substantial resemblance to those we have been skirting. It is made up of the usual cut edge-of the pampa, with a succession of light-colored cliffs, and with shingle beaches, flats, and foul ground along the shore. The highest land is in the neighborhood of Puerto Gallegos which, with the remainder of this bight of the Patagonian coast, was so much fre- quented by Mr. Beck during the course of the Brewster-Sanford Expedition. As intimated above, we are now in the zone of what we may broadly term the Magellanic marine avifauna, an assemblage of birds belonging to the southern tip of South America, to associated continental islands such as Staten Island and the Falklands, and in certain instances to islands of the Scotia Arc and the Antarctic Archipelago. Some of these species are strictly neritic in habit and are sedentary rather than migratory. Others, having their breeding grounds wholly or partly within the same region, are pelagic during a portion of the year, making an annual exodus that carries them, variously, a few degrees into lower latitudes, as far as the tropics, or even into remote oceans of the northern hemisphere. The breeding ranges of all species in the aggregation by no means coincide; they differ greatly, indeed, for reasons which may be either readily comprehensible or entirely obscure. In most instances such breeding ranges extend along a considerably greater length of coast on the Pacific side of the continent than they do in Atlantic Patagonia. Thus a composite of a number of characteristic examples would show a coastal distribution from near the southern end of the Golfo de San Jorge, in Patagonia, to close to the northern end of Chiloé Island in Chile or, roughly from latitude 47° S. on the Atlantic seacoast to 41° S. on the Pacific. The overbalance in favor of the Pacific within these latitudinal limits is probably due less to climate and to physical attributes of the ocean water than to better shelter among the maze of Magellanic islands, and to the mere territorial or areal advantage which is inseparable from the enormously greater length and complexity of coast line throughout southern 198 OCEANIC BIRDS OF SOUTH AMERICA Chile. Another controlling factor has to do with the enormous tidal rise along the Atlantic coast, which peters out very rapidly both inside the Strait of Magellan and along the coast toward the Strait of Le Maire. The significance of this in relation to waterfowl distribution is discussed below. Most of the Magellanic birds may be regarded as typically sub-antarctic. Some of them, however, will have to be called pan-antarctic because of their considerable extension toward polar land areas. Others, on the contrary, push their breeding ranges so greatly to northward on one or both coasts of South America—passing the tropic of Capricorn and even approaching the equator— that their precise distributional classification presents difficulties. These can be in part resolved through consideration of data presented in the preceding section on hydrology; the anomalies are most understandable along the Pacific coast, where the Humboldt Current carries a temperate environment far into the tropics. Doubtless the best method of expressing the different geographic relation- ships will be to divide a list of characteristic species into several categories, beginning with exclusively Magellanic and relatively sedentary salt-water forms and ending, as I have already done under a discussion of tropical birds, with brief references to species which are neither oceanic, nor even necessarily riparian, but which, nevertheless, reflect the zodgeographic circumstances it is desired to point out. Regarding the last group, it may be said that, whereas within the humid tropics numbers of continental lowland marsh birds reach the borders of the ocean, in the cooler belt of southern South America it is chiefly the species of the bare highlands that do so. It is hardly necessary to point out that the indicated distribution, from a more or less definite point on the Atlantic coast to a similar one on the Pacific, is rarely continuous. As suitable coastal habitats for any species are more often “spotty’’ than uniform, so are the ranges of the birds. Annotation of pertinent facts under the individual species will best serve our purpose. () SEDENTARY AND EXCLUSIVELY MAGELLANIC SALT-WATER BIRDS Divinc PetRELs Pelecanoides magellani. Coasts and inland waterways from Puerto Deseado on the Atlantic, southward to Cape Horn, and northward on the Pacific coast to Chiloé Island; approximately 48° S., Atlantic, to 42° S., Pacific. CoRMORANTS Phalacrocorax magellanicus. The Falkland Islands, and South American main- land and insular coasts from the Rio Santa Cruz, Argentina, to Chiloé Island; approximately 50° S., Atlantic, to 42° S., Pacific. Phalacrocorax albiventer. The Falkland Islands; South American coasts from Puerto San Julian, Argentina, to southern and western Tierra del Fuego; ap- proximately 49° S., Atlantic, southward to near Cape Horn. SOUTHERN ATLANTIC COAST 199 GEESE Chloéphaga hybrida. Southern Patagonia, on the Strait of Magellan, to Chiloé Island; approximately 52° S., Atlantic, to 42° S., Pacific. A distinct race (C. 4. malvinarum) inhabits the Falkland Islands. Like the South American Flightless Steamer Duck, discussed below, the Kelp Goose tends to avoid coasts characterized by large daily tidal amplitudes. This fact alone would limit its distribution on Atlantic shores. STEAMER Ducks Tachyeres pteneres (Flightless), T. patachonicus (Flying), and T. brachypterus (Falkland Flightless). The three forms of steamer ducks are thus classified in Part II of this book. T. brachypterus is confined to the Falkland Islands. The distribution of the mainland Flightless and Flying species largely coin- cides throughout a range extending from Puerto Deseado on the Atlantic, southward to Cape Horn, and northward on the Pacific coast to Corral. This is approximately 48° S., Atlantic, to 40° S., Pacific. The Flightless species (pteneres) avoids, however, coasts having a marked tidal rise and fall, with the result that it is normally absent from the whole eastern part of Tierra del Fuego and the Atlantic coast of Patagonia. Most of the steamer ducks in these dis- tricts represent the Flying species. To southward and westward, and through- out the Pacific extent of the range, the two are to be found together. OystTER-CATCHERS Haematopus leucopodus. The Falkland Islands; South American mainland and insular coasts from the Rio Chubut, Argentina, to Chiloé Island; approximately 43° S., Atlantic, to 42° S., Pacific. GuLts Leucophaeus scoresbii. The Falklands; Tierra del Fuego; probably breeds also northward well toward Chiloé Island on the Pacific side. While this gull is mainly sedentary, there is a more or less sporadic migratory movement that carries certain individuals northward as far as the Rio Gallegos, Argentina, and Mocha Island on the Pacific side. The species is not antarctic, and its inclusion within the South Shetland Island avifauna is probably based upon error. (2) MAGELLANIC MIGRATORY SALT-WATER BIRDS PENGUINS Spheniscus magellanicus. Breeds at the Falkland Islands and along South Ameri- can mainland and insular coasts from the Gulf of San Matias, Argentina, south- ward to Cape Horn and northward to the Bay of Arauco, Chile, or approximately from 42° S., Atlantic, to 37° S., Pacific. During the southern-hemisphere winter this penguin migrates northward to southern Brazil, sometimes as far as latitude 23° S., as previously noted. On the Pacific coast it wanders as far northward as central Chile (30° S.), overlapping the range of a closely related and representa- tive species, Spheniscus humboldti. . 200 OCEANIC BIRDS OF SOUTH AMERICA Skuas Catharacta skua chilensis. Breeds from southern Patagonia and eastern Tierra del Fuego to the Bay of Arauco, Chile, or approximately from 52° S., Atlantic, to 37° S., Pacific. Migrates northward at least to the coast of Brazil in the Atlantic, and into the northern hemisphere along the Pacific coast of America. Catharacta skua antarctica. Breeds at the Falkland Islands. (This is alleged also to be the resident race of Tristan da Cunha and Gough Island, but probably in error.) Migrates northward in the Atlantic to the edge of the tropics or beyond. The two Magellanic forms of the skua are arbitrarily placed in this category, rather than the next, despite the fact that they are both races of a bipolar species which has six or more subspecies. The Falkland and Chilean skuas are so distinct from the race inhabiting South Georgia, for example, as well as from one another, that each form may here be treated as though it were a full species. (3) SEDENTARY SALT-WATER BIRDS OF COMBINED MAGELLANIC AND HUMBOLDT CURRENT RANGE CorMORANTS Phalacrocorax gaimardi. From Puerto Deseado, Argentina, through the Fuegian waterways, and northward along the Chilean and Peruvian coasts at least to the Guafiape Islands; approximately 48° S., Atlantic, to 8° S., Pacific. OvysTER-CATCHERS Haematopus ater. The Falkland Islands; South American coasts from islands off the southern shore of Uruguay southward to Cape Horn, and northward along the Pacific to northern Peru; approximately 35° S., Atlantic, to 6° 30’S., Pacific. TERNS Sterna hirundinacea. The Falkland Islands; both coasts of South America from Cape Horn northward. On the Atlantic side it nests northward to the Gulf of La Plata (35° S.), and possibly to the vicinity of Cape Frio (23° S.). Along the Pacific coast it nests northward to San Gallan Island, Peru (14° S.), and prob- ably farther. In winter it wanders as far as Bahia, Brazil, and northern Peru. All antarctic records for this tern, such as the generally accepted legend that it breeds at the South Orkney and South Shetland Islands, etc., appear to be due to confusion of the species with Sterna vittata. (4) WIDELY DISTRIBUTED SUB-ANTARCTIC SALT-WATER BIRDS, WITH SOUTH AMERICAN BREEDING RANGES RESTRICTED TO THE MAGELLANIC DISTRICT PENGUINS Eudyptes crestatus. The Falkland Islands; easterly and southerly islands of the\Fuegian region. The westward progress of this penguin ends at the phyto- geographic limit of bare or grassy islands; it avoids forested or shrubby shores, SOUTHERN ATLANTIC COAST 201 such as characterize western Tierra del Fuego. The Jackass Penguin (Spheniscus magellanicus), on the other hand, occupies either bare, grassy, or wooded islands indifferently. ALBATROSSES Diomedza epomophora. Alleged to nest about Lake Cami or Fagnano, in Tierra del Fuego. SHEARWATERS Puffinus griseus. Nests at the Falkland Islands; islands of the Fuegian com- plex near Cape Horn; northward along the Pacific coast to Mocha Island, Chile (38° S.); and perhaps in the Andes of northern Chile. Divinc PETRELS Pelecanoides urinatrix. A race of this circumpolar species (P. u. berard) is resident at the Falklands and migrates as far as the coast of the Province of Buenos Aires. A less-well-known (or possibly the same) form occurs among the archipelagoes of southern Chile. (5) MAGELLANIC SALT-WATER BIRDS OF PAN-ANTARCTIC OR MORE EXTENSIVE BREEDING RANGE PENGUINS « Aptenodytes patagonica. South Georgia; Staten Island; formerly the Falkland Islands and various easterly outliers of Tierra del Fuego. ALBATROSSES Diomedea melanophris. South Georgia; the Falkland Islands; Ildefonso and other southerly Fuegian islets. Diomedea chrysostoma. South Georgia; the Falkland Islands; Diego Ramirez Islets, off Cape Horn. PETRELS Oceanites oceanicus. The South Shetland and South Orkney Islands; South Georgia; various Fuegian islets near Cape Horn. It appears, however, that the native Magellanic examples of this storm petrel represent a distinct and endemic race. CorMORANTS Phalacrocorax atriceps. South Shetland and South Orkney Islands; South American coasts from the Rio Santa Cruz, Argentina, to Corral, Chile; approxi- mately 50° S., Atlantic, to 40° S., Pacific. The exact specific and subspecific relationships of the pan-antarctic White- breasted Shag have not yet been satisfactorily worked out. Future study may make it necessary to change the status of the Magellanic representative, as here recorded. GuLLs Larus dominicanus. The South Shetlands, South Orkneys, and South Georgia; the Falkland Islands; South American insular and mainland coasts from Rio 202 OCEANIC BIRDS OF SOUTH AMERICA de Janeiro, Brazil, southward to Cape Horn, and northward to Lobos de Tierra Island, Peru; approximately 23° S., Atlantic, to 6° S., Pacific. In addition to all of the above, reference should be made to the Sheath-bill (Chionis alba), a bird of uncertain status, the eggs of which have not been found within the Magellanic region. The Sheath-bill nests at South Georgia and in many parts of West Antarctica. It is found, at times in relative abundance, at the Falklands, in Tierra del Fuego, and along the southerly Patagonian coast, and it may yet prove to be a permanent resident of some part of South America. The Bigiiaé Cormorant (Phalacrocorax olivaceus), which reaches the southern tip of South America from the north, is a sort of distributional antithesis of the Sheath-bill. Its ‘‘zoneless’’ type of range has been previously discussed, and it remains to say only that this curious cormorant is as much at home on the bold shores of Cape Horn as it is on Caribbean sandbars or in Brazilian lagoons. Finally, there are numerous birds which, farther north in South America, belong to the highlands or, at any rate, to the interior of the continent rather than the coast, but which in the Magellanic part of their ranges reach the edge of salt water. Among these are the following, the subspecific names, by which four or five of them have been distinguished from more northerly representatives, being omitted from the list. Little Grebe. Colymbus chilensis. Red-necked Grebe. Aechmophorus major. Night Heron. Nycticorax nycticorax. Black-necked Swan. Cygnus melancoriphus. Andean Goose. Chloéphaga melanoptera. Upland Goose. Chloéphaga picta. Ashy-headed Goose. Chloéphaga poliocephala. Coscoroba. Coscoroba coscoroba. Crested Duck. Anas specularioides. Spectacled Duck. Anas specularis. Gray Teal. Anas versicolor. Yellow-billed Teal. Anas flavirostris. South American Pintail. Anas spinicauda. Chiloé Widgeon. Mareca sibilatrix. Red Shoveller. Spatula platalea. Spurwing Plover. Belonopterus chilensis. Rufous-throated Plover. Oreophilus ruficollis. Pigmy Seed Snipe. Thinocorus rumicivorus. 7. Tue Atrtantic Sus-Antarctic IsLANDs Again we may leave the continent in order to consider the geogtaphy and avifauna of more or less remote islands. Staten Island is definitely part of the Fuegian district, and will be treated in connection with the mainland. South Georgia, although oceanic, is antarctic rather than sub-antarctic and is, more- over, a unit of the Scotia Arc, to which a separate section will be devoted. ATLANTIC SUB-ANTARCTIC ISLANDS 203 FALKLAND ISLANDS Miles 40 Kilometers ° 10 20 40 60 Heights in meters +o =. eS S a 2 Pebble I. , cS New I.3@ one 3. Beaver |. 4 £29) vin “ \Sea Lion IS ele 60° Bo Mego" 58° Fic. 28. This leaves, therefore, only the Falklands, the islands of the Tristan da Cunha group, and Gough or Diego Alvarez Island. The two latter are among the most isolated bodies of land in any ocean, while the Falklands lie entirely within the 200-meter contour of the Patagonian coastal platform. Nevertheless, the dis- tinctly oceanic elements in the bird population of the Falklands and of the Tristan area have many resemblances. a. Falkland Islands. The Falkland Islands lie on the continental shelf, about 418 kilometers east of southern Patagonia. With the exception of Beauchéne Island, in latitude 52° 55’S., they are included between the parallels of 51° and 52° 30’S., and the meridians of 57° 30’ and 61° 30’ W., their southern boundary being on a line with the eastern entrance of the Strait of Magellan. The group consists of more than one hundred islands and islets, but the bulk of the surface area of about 16,835 square kilometers is comprised within the two principal islands of East and West Falkland. These are separated by a strait or sound varying in width between 4 and 29 kilometers. East Falkland is in turn very nearly bisected into northeast and southwest halves by long fiords extending inland from the Atlantic and from Falkland Sound, respectively, a narrow central neck of land alone connecting the two main bodies. Ridges of rocky hills traverse broad tracts of turf without a tree, sloping down nearly everywhere to low rocky coasts. In the northern half of East Falkland, the chief range, known as Wickham Heights, runs east and west, culminating in Mt. Usborne, 684 meters high, near its western extremity. Practically all of the southern portion of the island, south of Choiseul and 204 OCEANIC BIRDS OF SOUTH AMERICA Grantham Sounds, is so low that it has very little visibility from the deck of a ship at a distance of 8 to 10 kilometers. West Falkland is of slightly greater average height, its peak, Mt. Adam, rising to 706 meters. Its hills run mostly north and south. Along the western coast of the group there are precipitous cliffs exposed to the fury of the almost incessant westerly winds. The Falkland Archipelago appears to differ in geological structure from the Patagonian mainland. The terrain resembles, however, the western part of Tierra del Fuego in showing evidence of very extensive glaciation. The innumer- able inlets and sounds indenting the shores of the islands are all of the deeply incised, ford type. The hilly surface is free from striking irregularities, although most of the uplands terminate in points and crests rather than in rounded or tabular formations. Low stony ridges, more or less devoid of vegetation, cross the land at irregular intervals and run in all directions, as if without plan. Even the lower-lying tracts are intersected by quartzite reefs which stand out in sharp contrast with the dark background of the peat-bogs. A peculiar feature in the geology of the islands is presented by the much-described “‘stone rivers,”’ which are made up of blocks and fragments of quartzite that have the appear- ance of flowing in narrow ribbons down the slopes of the hills. The lower land is made up of undulating plains, more or less resembling tundra, with winding brooks in the hollows and valleys, and with here and there boggy patches which show the dangerously warning color of bright green. Little seems to be known about relatively recent changes in the level of the land, but some of the older chronicles, such as that of Fanning (1838, 199), allege that entire skeletons of large whales have been found in elevated situations along the coast. The early navigators who saw the Falklands without landing, described them as being well wooded. These reports were owing to the islands being nearly covered with the most beautiful— as well as the most nutritious—grass in the world: ‘the golden glory of these islands’... . Patches of tussock-grass (Poa flabellata) look from a distance like small low groves of palm trees. From 200 to 300 shoots spring from one plant, which rises to a height of between 5 and 7 feet; the leaves are sometimes 6 feet long. They possess valuable feeding properties, being rich in pro- tein, and are eaten eagerly by horses, cattle, and sheep. A few inches of the bases of leaves, stem, and root are sweet and, according to Clayton, ‘when roasted eat like almonds.’ Ross says that two Americans who had wandered or deserted from their ship lived on West Falkland for fourteen months on tussock culm. The first French explorers thought it a kind of flag and called it glayeuls, though Bougainville recognized it as belonging to the graminae. Clayton, Penrose, and other Eng- lish called it Penguin Grass, because in it the penguins made their nests, ‘burrowing under ground like rabbits in their holes.’ Whether by accident or design it was frequently set on fire by the early settlers, with disastrous results to the birds and seals. . . . To be really happy it requires a peat bog soil with the sea-spray on it. As Governor Moody says: ‘Where the sea beats with the greatest violence and the spray is carried farthest, there the tussac grass thrives best on the soil it loves.’ Owing to the abundance of sheep, this grass has almost vanished from the East and Western Islands, though an attempt was made in 1922 to plant an area with it near Port Stanley. In some of the smaller islands—such as West Point—it still grows Juxuriantly, rising to a height of six feet, or more, the long leaves hanging gracefully down (Boyson, 1924, 276, 278). Such is a description of the most famous of Falkland plants, a grass which is, or was, of great importance in the ecology of the native birds, as it is also at ATLANTIC SUB-ANTARCTIC ISLANDS 205 certain antarctic islands, such as South Georgia. Another plant of interest is the peat-forming Empetrum rubrum, related to the crowberry of the northern continents. This is the often-mentioned ‘‘diddle-dee,’’ a word which, inci- dentally, was long ago transplanted to Nantucket, Massachusetts, if indeed the whalemen and sealers from that island did not originally introduce the name into the Falklands. Boyson (1924, 267) writes further: In a wind-swept country such as the Falklands one place seems to resemble another in outline and appearance, all being undulating grass-land and ‘diddle-dee.’ Looking down from a height the general aspect is grey. The rocks which, seen close, are streaked with red and gold lichens, become merged in grey distance; and grey seems the far-off, unquiet sea. The absence of trees— though, as will be seen, attempts are being made to introduce them—adds to the monotony. ‘Pas un arbre, pas un véritable arbrisseau ne vient rompre l’uniformité de ces vastes solitudes,’ wrote Dumont d’Urville in 1822. The blossoming shrubs bear no brilliant flowers to bring colour to the landscape, but delicately tinted blossoms, frail-seeming, many most sweetly scented, and almost hiding from view. It is the constant gales that prevent the growth of trees and large shrubs, except in the most sheltered places, and even there the tops are sheared off wherever they project beyond the windbreak. Consequently, the scenery in general resembles the northern moorlands of the British Isles. In the terse words of the ‘South American Pilot,’ Wind is the principal evil at the Falklands; a region more exposed to storms, both in summer and in winter, it would be difficult to mention. The winds are variable, seldom at rest while the sun is above the horizon, and at times very violent. During the summer a calm day is an extraor- dinary event. Generally speaking, the nights are less windy than the days, but neither by night nor by day nor at any season of the year are these islands exempt from sudden and severe squalls or from gales which blow heavily, though not usually lasting many hours. Bryce, after remarking that the “‘winds are so strong and incessant that everybody goes about stooping forward,"’ continues (1913, 314): I have seen many wild islands in many stormy seas, but never any inhabited spot that seemed so entirely desolate and solitary and featureless. There was nothing for the eye to dwell upon, no lake, no river, no mountain,—only scattered and shapeless hills,—a land without form or ex- pression, yet with a certain simple and primitive beauty in the colours of the yellow grass and grey-blue rocks, shining through clear air, with the sea-wind singing over them. The Falkland climate is of the sub-antarctic oceanic type, with marked equability throughout the year and frequent, though by no means heavy, pre- cipitation. There are showers, on the average, on 250 days in the year, and yet the mean rainfall amounts only to about 660 millimeters. October is the finest month and March the worst. The mean temperature of the two midsummer months (December and January) is 8.3° C., that of the two midwinter months 2.8° C. Frosts and snow are not uncommon even during the summer, but thick ice rarely forms on the ponds at any season and snow lies long on the ground only in the highlands. The surrounding ocean is always ice-free. The extent of overclouding is high, even though the air is often extraordinarily clear. Falk- land nights tend to be relatively brighter than the days, and the stars frequently shine with the brilliance of the tropics. 206 OCEANIC BIRDS OF SOUTH AMERICA All the island coasts are, of course, rich in intertidal life. Kelp and other sea- weeds pile up in enormous windrows along the beaches and, with flood and favorable gales, are likely to be rafted off to distant parts of the Atlantic. Cal- careous algae, sea-anemones, many crustaceans, mollusks and other inverte- brates abound everywhere in the shallow water, and the ocean surface is rich in nutrient material, as described in an earlier section of this book. Movements of the surface water are complicated around the archipelago because, in addition to the northward-setting flow of the Falkland current, there are strong tidal waves among some of the tortuous channels between the islets. In the neighborhood of the Jason Islands, which form the northwesterly extension of the whole group, the movement augments to a race with a speed of 10 to 11 kilometers per hour. The southerly shores of the archipelago, and the wide funnel at the southern end of Falkland Sound, are usually strewn with flotsam from the Fuegian region, which includes branches and trunks of trees as well as great masses of kelp and fragments of the wreckage of ships. Taste and industrial trends have not led the Falklanders to take to the sea to an extent transcending the necessary inter-island commerce, such as has been described in connection with the journeys of Mr. Beck. The inhabitants are, in fact, heirs of the gauchos, except that sheep rather than horned cattle are their present charges. As a constituent of island diet, mutton is known as “three sixty-fives."’ Bryce was struck by the anomalous absence of fishing- boats in the various ports. Despite a small total population, all but the very smallest islets of the archipelago are now inhabited at least by shepherds. So far as known, primitive man never reached the Falklands, and since there was supposedly but one native mammal—the endemic wild dog commonly known as the Falkland fox—the islands were a paradise of sub-antarctic bird life. The ‘‘fox’’ has been rare for a century and extinct since about 1875. Darwin (1839, 477) noted that the Falkland geese, which were originally fearless of man, were quite aware of the danger from the predacious dogs. Since the dis- appearance of the latter, in fact, there seems to have been a change in habit of certain waterfowl which were once confined to outlying islets, but which have now added the moorlands of the main islands to their breeding area. On the other hand, the settlement of the Falklands, and the successive introduction of horses, cattle, and sheep; the rivalry between sheep and such native birds as the Upland Goose (Chloéphaga picta) in the consumption of pasturage, which has led to the placing of a bounty upon the head of this bird; raids upon the penguin colonies for blubber-oil and eggs; the classification of hawks, caracaras, and skuas as pests; and other more or less inevitable human meddling, have con- spired to bring about an unfortunate reduction of the pristine bird life. The Falklands, being cold-temperate, have an avifauna which is somewhat analogous. in its components with that of high northern latitudes. While a number of the sub-antarctic species have no homologues in the northern hem1- sphere, the general aspect of Falkland bird life is not dissimilar from that of the fiorded coasts of Labrador and southern Greenland. Thus there is a profusion of gulls and terns (four species), a skua, several ducks and geese, numerous petrels ATLANTIC SUB-ANTARCTIC ISLANDS 207 and other Procellariiformes, and two native cormorants. The latter are both of sub-antarctic affinities, for the Bigiia Cormorant (Phalacrocorax olivaceus), which we have noted as occurring practically everywhere in America south- ward to the very island on which Cape Horn stands, has never reached the Falklands. This is curious, especially since the islands lie directly to leeward of waters inhabited by this cormorant and in a zone of powerful westerly winds. It goes to show that the Bigiia Cormorant, like the Brown Pelican, must be regarded as a species which closely hugs the continental seacoast, and which therefore could only fortuitously obtain a footing at relatively remote islands, as the pelican has done at the Galapagos. Visitors in different decades, even in different centuries, have obtained a notably uniform impression of Falkland bird life. Fanning (1838, 196) writes: ‘The various sea-fowl, resting upon the slight eminences, and brought in strong relief against the sky, ofttimes deceive the experienced eye of the mariner, by having their puny dimensions magnified in size to those of human form." Several recent voyagers have independently made the same observation, adding the turkey vulture to the list of sentinels that loom large on the mounds and headlands. The earliest voyagers also very quickly arrived at a gastronomic classification of Falkland birds, which time and taste have not modified. Clayton (1776, 104) and others give detailed notes on the relative edibility of the differ- ent species of geese and of the different kinds of penguin eggs, and name the season in which each may be taken to best advantage. Other mariners report the distance to which the various penguins and the King Shag may be seen off- shore from the Falkland colonies. The penguins travel so far that their presence may be discounted by shipmasters. Flocks of shags, however, will probably mean rocky coasts ‘‘within ten miles.” Bennett (1927, 259) makes the interesting statement that several species of Falkland birds lay smaller sets of eggs than their relatives in other parts of the world; both the resident oyster-catchers, for example, lay at most two eggs and not infrequently only one. The avifauna of the Falkland Islands comprises about 102 species (Bennett, 1926, 306), this figure including numerous ocean wanderers which never volun- tarily come ashore. The latter are chiefly petrels from the Antarctic or the north. These, together with visiting northern-hemisphere shore birds, regular migrants from Patagonia, casual or accidental species, and introduced birds, make up a third of the total. Bennett (1922, 256) reports extensive migrations of South American birds to the archipelago during years of continental drought, such as 1920. More than half of the 69 resident species are fresh-water or marsh birds and land birds, such as coots, plovers and their relatives, herons, Anatidae, vultures, hawks, owls, Cinclodes, flycatchers, wrens, thrushes, pipits, finches, and troupials. Chapman (1934, 6) lists six or seven endemic forms among the land birds, of which either two or three are to be regarded as full species. There remain the 29 waterfowl listed below. Not all of these are strictly salt-water forms, two of the geese, for instance, being grassland birds, and several others feeding more or less in ponds or elsewhere back from the coast. 208 OCEANIC BIRDS OF SOUTH AMERICA The list includes one endemic species, the Golden Grebe, Colymbus rolland, and at least two endemic subspecies; the status of some of the others is still to be determined. The King Penguin, and possibly one or more of the other birds, have recently become nearly or quite extinct at the Falklands. PENGUINS Aptenodytes patagonica Pygoscelis papua Eudyptes crestatus Spheniscus magellanicus GREBES Colymbus rolland Colymbus occipitalis ALBATROSSES, PETRELS, AND Divine PetrELs Diomedea melanophris Diomedea chrysostoma Macronectes giganteus Halobaena caerulea Pachyptila belcheri Procellaria aequinoctialis Puffinus griseus Garrodia nerets Pelecanoides urinatrix berard CoRMORANTS Phalacrocorax magellanicus Phalacrocorax albiventer GegsE AND Ducks Chloéphaga picta Chloéphaga rubidiceps Chloéphaga hybrida malvinarum Tachyeres brachypterus Tachyeres patachonicus OvysTER-CATCHERS Haematopus leucopodus Haematopus ater Skuas Catharacta skua antarctica GuLts AND TERNS Leucophaeus scoresbii Larus dominicanus Larus maculipennis Sterna hirundinacea The list should not be closed, moreover, without a reference to the Wattled Sheath-bill (Chionis alba), which has the appearance of a resident, even though it is less common in summer than in winter, and though its eggs have not with certainty been found in the Falklands (Bennett, 1926, 321). b. Gough Island. The islands of the Tristan da Cunha area constitute one of the most remote aggregations in any ocean. The southernmost and only completely isolated island in the region is Gough, or Diego Alvarez, the Portuguese name having priority by two centuries or more. It lies roughly 2500 kilometers west by south of the Cape of Good Hope, more than 3000 kilometers north by east of Cape Horn, and about 350 kilometers southeast by south of the nearest land, which is Nightingale Island of the Tristan trio. Gough Island extends northeast and southwest for a distance of 13 kilo- meters, its greatest breadth being about half its length. Of volcanic origin, it rises abruptly from the sea on every side, with wave-cut cliffs suggestive of Madeira or still more of St. Helena. These increase in height toward the north- ern end where they reach about 300 meters. The altitude of the highest peak ATLANTIC SUB-ANTARCTIC ISLANDS 209 North Pt GOUGH ISLAND Miles 0 ! 2 3 Kilometers Heights in meters Fic. 29. has been determined by Captain G. V. Douglas of the ‘Quest’ as 888 meters, instead of the 1300 meters previously reported. The island may be regarded as one of the pinnacles of the long mid-Atlantic ridge which makes its southern- most appearance above the ocean at Bouvet Island. Gough has been greatly eaten away by the sea, and the submarine contours, at least as far as the one marking a depth of 200 meters, have an outline very similar to that of the island itself. Off all coasts except the southwestern there are numerous exposed rocks and islets a short distance from shore. Several of these seem to be important as nesting grounds for certain terns and other sea birds to which the main island is unsuited. The largest are Lot’s Wife, the Cone, and Round Islet, off the northern end, and Penguin Islet off the eastern coast. Although Gough Island has never had permanent human inhabitants, it has been visited more or less infrequently for well over a hundred years, and numer- ous records of landings since 1820 have been carved on slabs of stones in the so- called ‘‘Glen’’ by whalemen, sealers, and ostensible diamond miners. The first scientific information about Gough came from collections and meteorological observations made by Captain George Comer, who has had a long association with the American Museum of Natural History, to which he presented the type 210 OCEANIC BIRDS OF SOUTH AMERICA of the Gough Island Rail (Porphyriornis comert) and other specimens. His journal, and notes on his collections, have been published by Verrill (1895, 430). More recent scientific reconnaissances are those of the ‘Scotia’ party in April, 1904, and of the antarctic explorers in Shackleton’s ‘Quest,’ made at the end of May and the beginning of June, 1922. The present notes are based in large part upon observations set down during these visits, as recorded respectively by Brown (1905, 430) and Wilkins (1923, 501). The naturalists of the ‘Scotia’ were surprised to find Gough Island beauti- fully clothed in green from the tideline almost to the summit, and to see streams tumbling everywhere from hanging valleys or pouring over sheer precipices into the sea. Above the brinks the ground slopes more gradually towards the summit, though a considerable proportion of the land appears nearly inaccessible. Throughout the island, steep radiating ridges are separated by narrow glens, many of which harbor good growths of trees. Landing is practically impossible on the western side, the water's edge of which, however, can be reached over- land, according.to Captain Comer’s diary. Excessive wave-erosion is testified to by the fact that most of the stream valleys have their courses truncated sharply by cliffs, so that the running water necessarily ends in a cascade. Some of the sea caves have evidently been left stranded at heights of from three to five meters above the present level of the ocean, indicating a relatively recent change in relation. Many mariners have stressed the difficulty or impossi- bility of effecting a landing at Gough Island except in quiet weather (Spiess, 1928, 98). Wilkins, however, went ashore from the ‘Quest’ on each of the four days between May 28 and June 1, which is close to the winter season. He states that the feat proved sometimes difficult but never dangerous to experienced men, and that by taking advantage of the shelving rocks at the foot of a cliff near the Glen, landing would be practicable in any weather. The southerly slopes of Gough Island, according to Brown, appear more fertile and luxuriant than those toward the northern end. The vegetation is, however, very dense nearly everywhere, presenting a serious obstacle to walk- ing about. This is partly due to the thick tufts of tussock grass (Spartina arundinacea), which is also the species of Tristan but is quite different from the tussock of the Falklands and South Georgia. Gnarled and stunted buckthorn trees (Phylica nitida) thrive from sea level to an altitude of almost 600 meters and grow even on exposed ridges. This is the species described nearly a century ago from Tristan da Cunha as Phylica arborea by du Petit Thouars, who said that it was the only tree growing on that island and that it reached a height of about 5 meters, with a trunk diameter of 15 centimeters or thereabouts. Later travellers have recorded boles up to 40 centimeters in diameter. The relation- ships of the genus are definitely with South Africa, where many other species are known. In a ravine south of the Glen, which lies near the middle of the east coast of Gough Island, members of the ‘Quest’ party discovered a grove of trees of a different sort, ‘growing as if planted in an orchard,”’ and reaching a height of 4 to 5 meters. These have been identified as Sophora tetraptera goughen- sis, a variety of leguminous tree peculiar to this island but very closely related ATLANTIC SUB-ANTARCTIC ISLANDS 211 to one native in the New Zealand region. A third arboreal plant, if it may be so called, is the low tree fern (Lomaria Boryana), which reaches a height of only a little more than a meter and is characteristic of the damper hollows and shel- tered glades throughout the island. The total Gough Island flora of about seventeen species of flowering plants, ten ferns and many other cryptogams is, as might be expected, very similar to that of the Tristan group, although it shows perhaps a stronger American element, as well as greater luxuriance. The coast is fringed with the usual border of kelp common to all sub-antarctic islands. Native insects include a number of species of beetles and flies. At the date of Comer’s visit, and apparently ever since, the island has been overrun with mice. These are generally assumed to be introduced house-mice. The only other mammals are the fur seals and sea-elephants which formerly hauled out in great numbers on all accessible shores. Data on the climate of Gough Island are chiefly in the form of general de- scriptions and impressions. Weather conditions are not very dissimilar from those at the Tristan group, concerning which a few figures are recorded below, but they are more severe, since Gough lies well within the Sub-Antarctic Zone of surface water, while Tristan is close to the convergence with the Sub-Tropical Zone. Spiess describes the torn gray clouds that seemed to whirl around the peak on July 1, and Wilkins not only repeatedly mentions the rain, but also states that snow covered the higher slopes on May 31 and June 1. Rainbows were a notable feature of the weather during the brief visit of the ‘Scotia.’ Comer’s residence at the island continued from August 23, 1888, until January 23, 1889. His journal of the weather and related phenomena is published in full by Verrill (1895, 472). It stresses the fog, gales, and excessive precipitation of this five-month period, and reports that a man of the party was frozen to death in the hills on October 19. Specifically, rain fell during six of the last nine days of August; on 19 days in September, with snow on three or four; rain or snow, or both, on 22 days in October; the same on 17 days in November, with many gales of wind, chiefly from a westerly quarter, but also with an increasing number of days characterized as “‘pleasant’’ in the entries; rain on 18 days in December, and on 15 days during the first 23 of January. Comer notes also the gradual warming of the ocean water at Gough Island with the advance of spring and summer. Between August 25 and 29 it ranged between 8.9° and 10.6° C., while a reading of the air-temperature on August 28 showed 11.1° C. Throughout September the ocean surface water remained between 10° and 10.6°; during October the range was from 10° to 11.1°; all the November entries record 11.1°; on December 24 it reached 12.2°, and on January G, 12:8"'C. The following record of landing, investigation, and personal experience at Gough Island is from the always illuminating and spirited pen of Sir Hubert Wilkins (1923, 504), who was naturalist of the ‘Quest’: Keeping close to the beach we followed the east coast, which presented a much more rugged and broken appearance than the steeper, higher, and less eroded western side. On the east side 212 OCEANIC BIRDS OF SOUTH AMERICA deep valleys ran steeply to the central ridge. One conspicuous peak, named the Monument, stood out boldly, and falsely assumed the appearance of the highest point. Twisted and straggling trees were silhouetted against the sky-line, and were noticed in rough perpendicular rows on the slopes facing the sea. This may be evidence of heavy snowslides, or perhaps of the distribution of seeds from the.trees at the higher levels. This theory is supported by the fact that the bigger and older trees are found at the top of the hills; on the other hand, it may be accounted for by the fact that the several parties who have lived on the island may have cut down for firewood all the bigger trees on the lower slopes. Jagged rocks footed the cliffs, which varied in height from 20 to 800 feet, and numerous caverns and caves were observed. The waves dashed on to the beach and into the caves with restless energy, and it was only on the rocks and sheltered coves that easy landings could be made. Here and there along the coast was a stretch of shingle beach, where the easier slopes approached the sea. It was on one of these and a few yards inland that two small huts were observed. One was the ordinary white man’s type built of weatherboard and roofed with iron; the other was a shanty, half underground, stone-walled, and thatched with the tussock grass of the island. The ‘Quest’ anchored about two hundred yards from the beach, and a landing was made in the early morning. Both huts were found to be in a very good state of repair, and the iron-roofed one proved a haven of refuge one night, when during a violent storm of hail and sleet our tent was literally blown from the ropes, leaving us exposed beneath the skeleton of ridge pole and guys. The wind, although not blowing a continuous hurricane, sweeps down the gullies and over the cliffs in terrific gusts at the rate of sometimes more than a hundred miles an hour. There was a quantity of stores, supplies, and household utensils found in the huts, and round about were strewn implements and apparatus commonly used by gold and diamond miners throughout the world. A ‘‘dolly”’ still on its rockers stood in position near a trickling stream of water that has been guided to the spot. Beneath this miniature waterfall has been placed a butt, now filled with water, and in this was the first evidence of the tragedy that happens to natural life when civilised man makes an entrance. The dark blue back of a Penguin floated level with the water; its head and legs hung limply in death, for its curiosity had led it into the tub, and, being unable to get out, it had slowly starved to death. Many other skeletons were noticed about the camp, terrible evidence of man’s destructiveness. . . . The weather during most of the visit to Gough Island was not favourable for observation generally, pouring rain, high winds, and low temperature not only making for physical discomfort, but also rendering hill-climbing difficult. The hills are mostly very steep, and it is only with the aid of the tussock grass, trees, and tree-ferns that one is able to scramble up their slippery slopes. The vegetation is almost tropical in its luxuriance, but at the time of our visit the ground was so wet the plants were easily pulled from the shallow earth and failed to give a firm foothold. The trees grow so close together on the easier slopes that it is impossible to walk around or under- neath them, and the branches are so intertwined that it is necessary in places to crawl through and over them for distances of a hundred yards or more. Clothes were torn to ribbons, and the pouring rain spoiled the appearance of most of the specimens collected. Wherever a sufficient depth of earth was discovered not overgrown with trees, it was honeycombed with holes made by the various species of Petrel that breed on the island. Very few of these birds remained during the daylight hours, but an occasional croaking betrayed a bird as the tramp of feet disturbed it in its resting-place. By this means several birds were located and added to the collection. Two Adamastor cinereus were discovered high up on the central ridge, and on the steep hillside above the hut four Broad-billed Prions (Pachyptila vittata keyteli) were secured. Many more were heard, but, as these birds have rather long and deep burrows dug beneath the tangled and tough tussock grass, it took some considerable time to dig them out. They were found in pairs in these holes, but no sign of a nest or the development of sexual organs showed a near approach to the breeding season. In the dusk of the evening the air was filled with cries of birds returning from their daily visits to the sea, and, although it was too dark to shoot in the ordinary way, a huge fire was lit and Major C. W. R. Carr, the aviator to the Expedition and an excellent shot, helped me to secure some birds as they flew through the firelit area. In this manner we secured several birds, but they were difficult to find in the high tussock grass when they fell out of range of the light. If they were ATLANTIC SUB-ANTARCTIC ISLANDS 213 left out all night they were found to have been almost entirely eaten by numerous mice that swarmed everywhere, even among our party as they lay in the tent at night. At the back of the huts and at several other points were Penguin rookeries, but they were deserted at the time of our visit, 1 June. They were invariably situated in the boulder-strewn narrow lowland between the high hills and the sea, and were partly hidden by the rank vegetation growing in the rich guano soil. A few old eggs and some skeletons were noticed here and there at the nesting-sites, but these were remarkably few, as was also the case at rookeries I have visited at South Georgia, South Shetlands, Graham's Land, and the Tristan da Cunha group. It is difficult to say what becomes of the old birds and the young ones that die from disease, for the numbers of skeletons in all rookeries visited would not amount to as many as one would expect to find in any one area as the result of death from natural causes. The two land birds of Gough Island are endemic. One is the rail mentioned above, belonging to a genus found elsewhere only in the Tristan da Cunha group. The other is a finch formerly associated with the Tristan genus Nesos- piza, but regarded by Lowe (1923, 512) as a peculiar genus and species, Rowettia goughensis. He holds that its affiliations are with Melanodera, the Yellow- browed Finch of southern South America and the Falkland Islands. Clarke (1913, 277) records 23 or 24 kinds of birds from Gough Island, but several of these represent merely pelagic transients collected over the adjacent waters. Below is a list of the oceanic species known, or assumed on reasonable evidence, to breed at the island. Future research may extend it considerably. PENGUINS Pterodroma mollis Eudyptes crestatus Fregetta grallaria (2) Garrodia nereis ALBATROSSES AND PETRELS f : } Pelecanoides urinatrix dacunhae Diomedea exulans dabbenena Diomedea chlororhynchos Phoebetria fusca Sxuas, GuLtts AND TERNS Macronectes giganteus Catharacta skua, subspecies? Pachyprila forsteri Larus dominicanus Adamastor cinereus Sterna vittata Puffinus assimilis elegans Anoiis stolidus (2) While the ‘Scotia’ was lying off Gough Island from April 21-23, 1904, all of the birds in this list except the Kelp Gull and the Noddy Tern were captured or observed. In addition, as recorded by Wilton, Pirie, and Brown (1908, 75), the Scottish naturalists collected the following four species which are not known to nest at the island and which belong, in fact, toa more southerly breeding zone. Diomedea melanophris Procellaria aequinoctialis Phoebetria palpebrata Oceanites oceanicus The three albatrosses breeding at Gough Island are all “‘low”’ sub-antarctic forms, even the resident Wandering Albatross being of a race quite distinct from that inhabiting South Georgia and other islands lying in the Antarctic Zone of surface water. The lesser albatross (Diomedea chlororhynchos) is also different from species nesting at the Falklands or elsewhere in the higher sub-antarctic 214 OCEANIC BIRDS OF SOUTH AMERICA latitudes. In the presence of such species as Pterodroma mollis and Anois stolidus we can discern even the suggestion of an approach toward the Sub-Tropical Zone. It is not established that the Noddy nests at Gough Island, although it has been reported in winter as well as in summer, and it breeds no farther away than the Tristan group. Comparing the whole marine avifauna of Gough with that of the Falklands, we find perhaps six species common to each, but an even larger number of related forms are distinct either as to species or subspecies. Furthermore, Gough lacks entirely the cormorants, grebes, ducks, geese, swans, and shore birds represented at the Falklands, and boasts only a rail and a finch as reminders of remote zodgeographic bonds with the continents. Wilkins observed at Gough Island the definite predilection that insular sea birds appear to have for windward coasts throughout the belt of strong wester- lies. During the entire stay of his vessel the notable concentration of albatrosses, petrels, and terns was always to be found over waters to westward of the island, scarcely any birds being observable in the lee. The phenomenon reminded him that the great albatrosses at South Georgia and other islands of the southern Oceans seem to show a cognate preference in making their nests chiefly in west- erly situations, where they are most exposed to wind and weather but as a con- sequence enjoy optimum advantages in the relatively difficult matter of taking flight. Just why such birds should still find windward coasts attractive after they are once in the air is less clear; possibly there are advantages connected with a supply of food at the surface. c. Tristan da Cunha. The islands of the Tristan group are referred to so frequently in the biogra- phies of the Rockhopper Penguin and of several albatrosses and other birds that a brief description will here suffice. The three members form a triangle within easy sight of one another, the peak of Tristan, which is in latitude 37° 5’ 50’ S., longitude 12° 16’ 40’’ W., being, indeed, visible in clear weather for well over 100 kilometers. As with Gough, their pedestals are on the mid- Atlantic ridge but there are soundings of 2000 meters or more between them. The closest continental point is the Cape of Good Hope, distant about 2500 kilometers; St. Helena and South Trinidad, the nearest islands except Gough, lie more than 2000 kilometers beyond the northerly horizon. Tristan da Cunha itself is 11 kilometers in diameter, nearly circular, and with the profile of a truncated cone projecting at an angle of forty-five degrees. The height of the peak is 2329 meters, and in the extinct terminal crater is a lake around which albatrosses were accustomed to nest at least until after the date of the ‘Challenger’s’ visit in 1873. The coast is mainly precipitous, with cliffs from 300 to 600 meters, broken here and there by steep gulches ending in shingle beaches. Along the northwesterly side, however, an extensive sloping grassy shelf, upon which the settlement is established, stretches for about 4 kilometers in front of the abruptly rising terrain, and terminates in Herald Point. The coasts of Tristan da Cunha have a scanty covering of green, derived mainly from grasses, sedges, mosses, ferns, the crowberry (the same as the ATLANTIC SUB-ANTARCTIC ISLANDS 215, TRISTAN DA CUNHA ISLANDS Miles Sédinburgh i) 5 10 Settlement Kilometers 5 we . . 2329 Heights in meters ae Tristan I. Daley Pt South Pt North Pt : “_Inaccessible I. WestPt( | Sél East Pt South Hill & Stoltenhoff I. > Middle I. C8, ~~ Nightingale |. 293 30" Fic. 30. ““diddle-dee’’ of the Falklands), and the darker patches of buckthorn trees. Many plants are of northern-hemisphere affinities, such as a bright green dock (Rumex), which grows under the bluffs and along the numerous watercourses. Dotted among the other herbage are the bluish green tufts of the tall reed-like grass (Spartina arundinacea) which is the so-called tussock peculiar to the Tristan group, Gough, and Amsterdam Island in the Indian Ocean. In many respects, this is the most important plant of Tristan and its neighbors. Moseley (1879, 117) writes of it as follows: In Penguin rookeries the grass covers wide tracts with a dense growth like that of a field of standing corn, but denser and higher, the grass reaching high over one’s head. . . . Here there is a sort of mutual-benefit-alliance between the penguins and the tussock. The millions of penguins sheltering and nesting amongst the grass, saturate the soil on which it grows, with the strongest manure, and the grass thus stimulated grows high and thick, and shelters the birds from wind and rain, and animals, such as the predatory gulls. The lower slopes of the mountain are covered with brush and ferns, including a small tree fern. Formerly this might appropriately have been called the forest 216 OCEANIC BIRDS OF SOUTH AMERICA belt of the island, but destruction of the woody vegetation progressed very rapidly after the beginning of colonization in 1817. By the time of the “Chal- lenger’s’ visit, good-sized examples of the one large tree, the native buckthorn (Phylica nitida), were to be found only in remote gorges of the mountains (Thomson, 1878, 34). The constant heavy gales do not permit the tree to grow erect; the trunk is usually procumbent at its origin for several feet, and then rises again often at a right-angle. It is always more or less twisted or gnarled. In sheltered places, as under the cliffs on the north-east of Inaccessible Island, the tree is as high as 25 feet, but it is not nearly so high on the summit of the island, though the trunks are said to reach a length of 30 feet or more. The largest trunk I saw was about one foot in diameter, but they are said to grow 18 inches (Moseley, 1879, 112). Giant kelp grows sparingly in shallow water around all of the islands. Inaccessible and Nightingale Islands lie southwestward from Tristan at a distance of 32 kilometers, and:about 18 from each other. The first of these is diamond-shaped, approximately 5 kilometers in length between its east and west points. The highest peak, situated near the western end, is 561 meters in altitude, and is said to enclose an old crater containing water, like that of Tristan. From this hill the land slopes irregularly, terminating in lofty cliffs except on the northerly coasts where there is a foreshore with negotiable landing places, dense growths of tussock, and even small copses of the buckthorn trees covered with a drapery of lichens. In addition to the main peak, there are two lesser rocky cones on Inaccessible. Wilkins landed at this island on May 22, 1922, and, by making full use of the tough tussock, he and other members of the ‘Quest’ party hauled themselves to the plateau after a stiff climb from the beach on the eastern side. Nightingale is really a group of three islands lying along a north-south line. Nightingale itself is the southernmost, and is somewhat less than 2 kilometers in length from east to west, with two peaks, the higher of which is upwards of 300 meters in altitude, appearing of conical form in certain aspects. Most of the coasts are precipitous and have been carved by the sea into deep recesses and caves. The explorers from the ‘Quest’ crossed the island on May 21, 1922, and found that the way to the summit could be forced only with great difficulty through the high tussock grass. Many petrels were found in their burrows, and three examples of the Greater Shearwater (Pxffinus gravis) were collected at a season when most examples of the species had previously been supposed to be on their migration toward the northern hemisphere. From the anchorage, at night, innumerable calls were heard of petrels returning from sea. A large number of Kelp Gulls was also observed at the island. To northward of Nightingale, beyond a narrow strait which is blocked by rock fragments, lies the relatively low Middle Islet; and 1500 meters to north- ward of the latter is Stoltenhoff Islet, named for the marooned Germans who were taken from Inaccessible by the ‘Challenger.’ This island is precipitous but flat-topped, and is broken into three parts by narrow chasms. The Tristan da Cunha group has a wet and stormy climate, with frequent precipitation during at least nine months of the year. There are, unfortunately, ATLANTIC SUB-ANTARCTIC ISLANDS 217 no comprehensive meteorological figures, but Wilkins (1923, 502) gives the following records relating to the year 1908: rainfall, 1646 millimeters; sun all day, 126 days; sun part of day, 232 days; sun unrecorded, 7 days; extremes of temperature 3.3°-25.4° C. Hail, snow, and ice are very rare near sea level, though between June and September the summit of the mountain is usually snow-covered. The finest season is from mid-December to March, January and February especially being usually bright and sunny months, with light breezes. Shortly before the end of the most unpleasant season, namely in mid-September, the majority of the oceanic birds which breed at the islands ‘‘begin to muster in couples on the sides of the mountains, and fill the air with the sound of their noisy courtship’’ (Earle, 1832, 366). Moseley found that the temperature of the fresh-water runnels on Tristan was 10° C., and he inferred that the lower levels of the island were constantly cooled by the descent of such water from the snow fields far above. Rollers, or great oceanic waves such as have been described for Ascension and St. Helena, sometimes reach Tristan da Cunha and pile high against the rocks. Early in the nineteenth century one such cast the British warship ‘Julia’ ashore on the island with heavy loss of life. Of the domestic animals introduced into the Tristan islands, pigs, goats, rabbits, cats, and rats have successively become feral. The first three were sub- sequently exterminated at Tristan itself, two of them by the human inhabitants, and the rabbits and cats (!) by rats, which in turn have unquestionably been responsible for the disappearance of many native birds. An account of the accidental introduction of rats into Tristan, and of the subsequent effects, is given on page 481. The dogs of the islanders are also known to have been very destructive to ground-nesting petrels. At Inaccessible Island, wild hogs, which feed largely upon sea birds and their eggs, have existed for a great many years (Thomson, 1878, 153). One entire penguin colony had been wiped out by these animals even before the time of the ‘Challenger’ expedition. According to Gordon (in Mathews, 1932, 21), ‘‘seemingly there are no rats yet on either Inaccessible or Nightingale Islands, and long may it continue so.”’ To ornithologists familiar with phenomena of insular evolution, it will not seem surprising that each of the three islands of Tristan da Cunha harbors endemic races of land birds, and that Inaccessible is the home of a unique rail (Atlantisia rogersi), which is one of the most extraordinary flightless birds in the world. Its origin, and the manner in which its ancestors may have reached this speck in the wide Atlantic, pose many obscure problems. The remaining land birds are the Tristan Coot (Porphyriornis nesiotis), with a cousin at Gough Island; a species of thrush (Nesocichla), with distinct subspecies on Tristan and Inaccessible; and two species of buntings (Nesospiza), one of which is repre- sented by separate races on Tristan and Nightingale Island; the second, Nesospiza wilkinsi, is a giant form peculiar to Nightingale. The land birds of Tristan da Cunha itself have all become extinct, and many of the sea fowl have been driven to residual sanctuary at the two lesser islands, where man has long been a devastating visitor but is not yet a permanent curse. 218 OCEANIC BIRDS OF SOUTH AMERICA The Tristan da Cunha group shares all the native sea birds of Gough Island save the Giant Fulmar (Macronectes). The latter has been included among the resident species by Peters (1931, 46) and several preceding writers, but apparently in error. In no part of the world, I infer, does its zonal nesting range extend northward quite to the Sub-Tropical Convergence; Gough Island seems, indeed, to represent the closest approach. Tristan, on the other hand, has seven or eight residents not yet recorded from Gough, but it would be rash to make a hard-and-fast comparison before the avifauna of Gough is much better known than at present. The ultimate residue in favor of Tristan will doubtless include some of the following species, especially such as have at least a slight sub-tropical stamp. Puffinus gravis Pelagodroma marina Pterodroma macroptera Fregetta grallaria Pterodroma incerta Anous minutus Pterodroma externa One would expect Prerodroma macroptera, in particular, to nest also at Gough. The status and ranges of the various Mother Carey’s chickens of the genus Fregetta are still very uncertain, despite recent discussion by Mathews (1928, 7; 1932, 24). 8. Tue Scotia Arc The narrowest gap between south polar lands and the rest of the world is that separating the Antarctic Archipelago from the southern tip of South Amer- ica. Even here the barrier is a great one from a biotic point of view, because Drake Strait is nowhere less than 700 kilometers in breadth, while the abyss of its waters descends to 4000 meters. To eastward and northward of the Antarctic Archipelago a garland of islands curves like a sickle toward Cape Horn, presumably connecting the lines of the Patagonian Andes with those of the Antarctandes. From the South American end, Tierra del Fuego, Staten Island, the Burdwood Bank, Shag Rocks, South Georgia, Clerke Rocks, the South Sandwich group, the South Orkneys, Clarence and Elephant Islands, the South Shetlands, and the islands once collectively called “‘Graham Land,"’ complete the loop, which is known as the Scotia Arc. Echo-soundings show that most of the submerged portions can still be traced in the relief of the sea bottom. At the bend of the bridge lies the South Sandwich chain, closely paralleled to eastward by a narrow trench with contours going 8000 meters beneath the surface. All the evidence seems to indicate a protrusion of circum-Pacific tectonic structures into the heart of the South Atlantic, repeating the condition of similar arcs in the Caribbean region. The contrast between the islands of the Scotia Arc and any of the insular or continental sub-antarctic districts we have thus far discussed is a profound one. Most of these American antarctic islands, as pointed out by Holtedahl (1931, 401), are now in the climatic state of Norway during the Ice Age. The more southerly members are covered in ]arge measure by an antarctic ice-mantle type of glacier, 7. e. the sort produced when ice covers the greater part of the land THE -SCOTTAYADG 219 mass down to sea level, and yet is of no great thickness, the undulations of the terrain being still registered on the surface. Within historic time considerable direct evidence has been noted that this ice-domination is on the wane, and that the action of the foreland glaciers is gradually subsiding. Nevertheless, the environmental difference between such a region as Tierra del Fuego or the Falklands and the islands of the Scoria Arc, all of which lie within the Antarc- tic Zone of surface water, is very great. The more southerly of these islands are, indeed, hardly removed from extreme polar conditions. Even those farthest from the antarctic circle, like South Georgia and the northern members of the South Sandwich group, which because of milder winters, a higher temperature range, a greater amount of ice-free surface, and more abundant precipitation, have a perceptible plant cover, are still characterized by a greater paucity of life than anything known in continental South America, or even in truly polar lands of the northern hemisphere. The source and effect of the antarctic oceanic climate have perhaps never been stated more clearly than by Darwin (1839, 274) when pointing out the significant differences between the productions of the antarctic islands and of corresponding latitudes in the northern hemisphere: On the northern continents, the winter is rendered excessively cold by the radiation from a large area of land into a clear sky, nor is it moderated by the warmth-bringing currents of the sea; the short summer, on the other hand, is hot. In the Southern Ocean the winter is not so excessively cold, but the summer is far less hot, for the clouded sky seldom allows the sun to warm the ocean, itself 4 bad absorbent of heat; and hence the mean temperature of the year, which regulates the zone of perpetually congealed undersoil, is low. Notwithstanding these considerations, the long extension of the Antarctic Archipelago toward relatively mild winds in the belt of westerlies has endowed this section of Antarctica with at least relatively favorable life conditions, making it on the whole the richest part of the south polar regions. If we assume that all its life originally entered from the north, we may say that the peninsuloid protrusion of a chain of high islands from the antarctic continent toward the South Shetlands and Joinville Island has enabled no fewer than ten species of birds to gain a local foothold in Antarctica. Furthermore, stray con- tinental birds appear to enter this region with extraordinary frequency. A North American migrant sandpiper (Bartramia longicauda), and the South American Pintail Duck (Anas spinicauda), have been taken at Deception Island, latitude 63° S. (Bennett, 1927, 259), while in 1916 and 1917, after a prolonged drought in Argentina, so many examples of a central South American duck (Oxyura vittata) migrated to the South Shetland Islands that remains of these birds were commonly found for several years thereafter (Bennett, 1920, 30). Such casual happenings are suggestive in connection with the permanent immi- gration into the Antarctic of terns, skuas, sheath-bills, etc. a. South Georgia. If we start eastward from Staten Island, the first considerable land mass encountered in the Scotia Arc is South Georgia, lying between latitudes 54° and 55° S., and longitudes 35° 50’ and 38° 15’ W., and approximately 4000 220 OCEANIC BIRDS OF SOUTH AMERICA Newark B. Annenkov Boe oe Undine South Hore, Boy C.Charlotte * “WIS : Gold Harb. Piokeregill) le Novosilski 8. : -Filchner SOUTH GEORGIA ISLAND C. Disappointment * Miles * 10 20 Kilometers 6 0 20 30 Heights in meters Fic. 31. square kilometers in area. Its length is 160 and its greatest breadth about 30 kilometers. The trend of the island is southeast-northwest and it is crossed in the same direction by a mountain range with folds overthrust toward the north. The coast is deeply fiorded, the drowned valleys filled by glaciers as mighty as those of Spitsbergen, and snow-capped peaks rise to elevations of well toward 3000 meters. From shipboard in clear weather the island presents one of the world’s most glorious spectacles—like the Alps in mid-ocean. The present snowline lies at an altitude of about 500 to 600 meters which, it should be noted, is lower than the treeline in the same latitude on Tierra del Fuego. Along the northeastern side of South Georgia a wind likened to the Alpine foehn frequently sweeps down the mountains and tends to prevent the accumu- lation of snow on the lower slopes, even in winter. Such ‘‘williwaws’’ are sometimes strong enough to roll stones along level ground; they are usually accompanied by a rise in temperature, owing to pressure-heating of the descend- ing air. The proximity of Weddell Sea, with its tendency to project cold weather northwards, produces generally low temperatures, the annual mean being 1.4° C., or fully 5° lower than that of Tierra del Fuego. Moreover, the low-pressure nucleus of Weddell Sea exerts the same influence as at the South Orkneys, so T PEC OA! As. ABC 224 that the temperature curves of these islands and South Georgia parallel each other. The temperature amplitude is low because of the oceanic influences, the mean for June, the coolest month, being —2.0° C., that for February, the warmest, 4.9° C. The recorded absolute minimum and maximum temperatures are —12.3° C. (July) and 20.7° C. (February). The mean temperature of the seawater in Royal Bay during 1882-1883 was 1.63° C. Precipitation is relatively abundant, an annual average of about 1300 millimeters falling even on the less- exposed northeastern side of the island. The monthly total is least in October and highest in April. Fullness of the glacial streams depends rather upon the melting than the falling of snow; the torrent season begins in November and doubtless has an important local effect upon the enrichment of the sea. Although South Georgia is often spoken of as sub-antarctic, as by Mecking (1928, 319), it is in an oceanographic sense an outer antarctic island, more or less surrounded by ice during several months of each year. South Georgian vegetation reflects the climatic conditions. There are but 19 species of vascular plants, of which the most pretentious is a grass. Of cryptogams about 200 species have been recorded. The affinities of the flora are in the main with South America, to a lesser extent with easterly islands toward New Zealand. The flora represents a transition between the Falkland type and that of the Antarctic Archipelago. The most abundant and conspicuous plant is the virile tussock grass (Poa flabellata), which covers much of the low ground and on favorable northern exposures straggles up to altitudes of three or four hundred meters. As at the Falklands, this grass loves the vicinity of the sea and thrives particularly well on the low-lying islets off the coast, locali- ties that never suffer from the detriment of an ice cap, and on which the fallen snow is always exposed to the denuding violence of the gales. The other grasses of South Georgia are inconspicuous. The plant next in prominence to the tussock is a rosaceous herb (Acaena adscendens), common also to the southern islands of the Indian Ocean. The round red-flowering heads of Acaena blossom out attractively on the moraines and lower hillsides all through the South Georgian summer, often pushing through beds of lingering snow. All but three or four of the other phanerogams must be diligently sought to be seen at all. Then there are four species of modest ferns of which two are rare, and a variety of liverworts, mosses, lichens, and algae. An orange-colored lichen (Placodium) forms brilliant patches on the ledges along the coast, and a shaggy gray Usnea beards many of the rocks. About the first of February, small red toadstools (Marasmius) spring up between the tussock hummocks on the higher beaches, and for the remainder of the brief southern summer add one homelike feature to the cold ground. Girdling the shore, and dampening the action of surf along many stretches, are dense growths of the giant kelp (Macro- cystis), the longest of sea plants which harbors among its branches an aggrega- tion of living creatures more abundant than any inhabiting forests of the upper world. When the ocean is calm, and the sun peeps through the clouds, sending a shaft of light down into the brown obscurity of the kelp beds, the water may be seen to beliterally filled with minute beings which constitute part of the plankton. rh) OCEANIC BIRDS OF SOUTH AMERICA South Georgia has the richest insect fauna of any American island lying wholly within the zone of antarctic surface water. There are five species of beetles, four of Diptera, and five of Collembola, not counting a considerable number of parasitic forms. Mecking is over-optimistic when he states that South Georgia seems to be not without prospects of success as a sheep country. As a matter of fact, intro- duced sheep and European rabbits, both of which thrive at Kerguelen Island, have failed to survive at South Georgia. The severity of the climate seems to be just beyond the threshold of conditions compatible with their existence. On the other hand, feral horses and reindeer have found the climate and the rich pastur- age quite to their liking, while rats have been acclimated since the early days of fur seal hunting, probably since about 1800. Still another introduced animal is the much-persecuted Upland Goose (Chiséphaga picta) of the Falkland Islands. A few pairs were introduced from the Falklands in 1910, and have since thriven and multiplied in the grasslands about the several fiords of Cumberland Bay, where they should be forever untroubled by the rivalry of sheep (Murphy, 1916, 276). The native birds of South Georgia have been studied during two American Museum expeditions, as described in an earlier section (pp. 25,26). They include two endemic species, namely, a pipit CAnthus antarcticus) and a teal (Anas georgica). Both of these are representatives of species inhabiting southern South America, directly to windward. The teal is, in fact, a close relative of the Brown Pintail (Anas spinicauda). No other species at the island is more than racially distinct from its nearest of kin. With the exception of the teal and the pipit, the avifauna is thoroughly characteristic of the Antarctic Zone. It lacks, indeed, only four birds found in the highest group of latitudes to which the sea extends, namely, the Emperor and Adélie Penguins, the south polar form of the skua, and the Antarctic Petrel (Thalassoica). The complete list of breeding sea birds of South Georgia is as follows: PENGUINS Aptenodytes patagonica Pygoscelis antarctica Pygoscelis papua Eudyptes chrysolophus ALBATROSSES AND PETRELS CorMORANTS Diomedea exulans exulans Phalacrocorax atriceps georgianus Diomedea melanophris Diomedea chrysostoma Ducks Phoebetria palpebrata Anas georgica Macronectes giganteus Daption capensis (probably) SHEATH-BILLS Procellaria aequinoctialts Chionis alba Pagodroma nivea Occanites oceanicus Sxuas, GuLLs AND TERNS Fregetta tropica Catharacta skua lonnbergi Garrodia nerets Larus dominicanus Pelecanotdes georgicus Sterna vittata georgiae PRESS COERTA ARE The breeding record of the Snow Petrel (Pagodroma) perhaps needs con- firmation. Moreover, if this species nests at South Georgia, it is hardly likely that Priocella antarctica does not. It should be remembered that the mighty steeps and headlands of the southwesterly, windward, and colder coast of the island are still little known. Probably a search of suitable localities near the southern extremity would disclose Daption, Pagodroma, and Priocella as common resident petrels. b. South Sandwich Group. The South Sandwich Islands, which form an arcuate chain at the eastern apex of the Scotia Arc, may still be called one of the least-known archipelagoes in the world. Discov- ered by Cook, in January, 1775, they were mistaken for promontories of an antarctic mainland, but von Bel- lingshausen, in December, 1819, dis- tinguished several of the islands, observed their volcanic, steaming character, and even pointed out that the penguins made use of land kept clear of snow and ice by the warm vapors. He also mentioned particu- larly the large numbers of sea birds of other sorts seen about the group and over the neighboring fields of floating ice (von Bellingshausen, 1902, 56). Morrell (1832, 65) later visited sev- eral of the islands in the ‘Wasp,’ and recorded the presence of albatrosses, Giant Fulmars, skuas, and Sheath- bills, in addition to at least one bird conjured out of his imagination. Coming to modern times, the ‘Scotia’ passed the southerly islands in 1903, and Larsen made important Jandings upon a number in 1908. Dur- 223 28° 20 | | SOUTH SANDWICH ISLANDS © Zavodovskil Traverse Islands C>Visokoi I. Ss Candlemas IS =r Vindication I. o Wealopes ‘ Kilometers § 40 20 30 40 Heights in meters Montagu |. Forster's Passoge Southern Thule ¢ Bellingshausen I. I Wl. 2 ! hule or Morre LG eock I. Fic. 32. 224 OCEANIC BIRDS OF SOUTH AMERICA ing the spring of 1911, the ‘Deutschland’ cruised along the chain, and Filchner (1922, 111) has written of the abundance of penguins, Cape Pigeons, and Snow Petrels. The ‘Quest’ made a brief reconnaissance of the northerly members dur- ing January, 1922, Wilkins (1923, 490) subsequently publishing some excellent observations on the bird life, most of which are quoted in my biographies. The ‘Meteor’ later passed to northward and eastward of the group in the middle of the summer period, and plumbed, by means of echo-soundings, the adjacent deep, which shows an extreme of 8060 meters (Spiess, 1928, 199). The explorations listed above all culminate in the thorough reconnaissance made by R.R.S. ‘Discovery II’ at the end of February, 1930 (Kemp and Nelson, 1931, 133). In the course of this notable cruise the British research vessel steamed at slow speed around each of the islands. The position of each was fixed by astronomical observation, and continual soundings were carried out. Good use was made of the camera, and provisional maps were prepared of the eleven islands and their outliers. Twenty days were spent in the survey. Tem- peratures proved low, even in this midsummer period, and for one week the thermometer never rose above the freezing-point. The group lies between latitudes 56° 18’ and 59° 28’ S., and between longi- tudes 26° 14’ and 28°11’ W. The northernmost island is about 535 kilometers east 1% south from South Georgia. The chain is over 350 kilometers long, and concave toward the west. The largest member is Montagu, circumference about 44 kilometers, and the smallest Leskov, with a circumference of about 3 kilo- meters. All of the islands are volcanic, several of them still being in the stage of simple cones. Five have definite present activity, emitting vapors of internal warmth, while three others, namely, Bristol, Cook, and Thule, are very heavily glaciated, without sign of warmth and with an ice cap at sea level attaining an average thickness of 60 meters. Candlemas Island has an active crater on the southern side of the summit, and several lesser cones that emit steam and sul- phurous fumes. Penguins avoid this part of the island, but elsewhere they occupy the ground in large numbers wherever the absence of glaciers and cliffs permits them to gain the shore and establish themselves. The ‘Discovery II’ party specifically noted penguin rookeries upon four of the other islands as well. The tussock grass of South Georgia apparently grows nowhere in the South Sandwich group. Giant kelp is also completely lacking in the waters round about. (Kelp occurs at the South Orkneys, though much less abundantly than at South Georgia, while it is missing at the South Shetlands.) Many icebergs were found stranded near the shores of the South Sandwich Islands. Between the group and South Georgia a tabular berg with an estimated length of over 100 kilometers was encountered; the ‘Discovery II’ rode out a full gale of wind under the lee it afforded. During the winter, of course, the South Sandwich Islands are completely beset by pack-ice from Weddell Sea. The ornithological status of the archipelago may best be summarized in the words of the ‘Discovery II’ naturalists: Birds are numerous on the islands. Penguins are especially abundant and have built their rookeries in almost every suitable place. They will not nest on the glacier itself, nor, doubtless THE ' SCOTIA’ ARC 225 because of the danger of falling ice, on the beaches beneath it. In consequence they find much difficulty in establishing themselves on the more heavily glaciated and precipitous islands; on some, such as Thule and Bristol, there are only one or two small headlands which they can occupy, and they appear to be altogether absent from Cook and Leskov. On islands where the warmth of the rock has melted much of the snow they are particularly abundant, and on Bellingshausen it was noticed that small groups had climbed the steep slopes of the cone, ascending almost to the edge of the crater and to a height of at least 450 ft. above sea-level. The Antarctic or Ringed Penguin (Pygoscelis antarctica) is much the commonest species, and Dr. Marshall, who landed near a small rookery on Thule Island on March 7, tells us that he found the young birds shedding their down coats. Gentoo and King penguins (Pygoscelis papua and Aptenodytes patagonica) were seen on Zavodovski and Saunders Islands. They probably occur on other islands also, but even with powerful glasses it was difficult to identify the species of penguin from on board ship. Those frequently seen on icebergs were invariably P. antarctica. Capt. J. Brown, who landed a party on Zavodovski in 1830, reported that his men found five different species of penguin, but he does not give their names. On this island Larsen found the Macaroni penguin (Exdyptes chrysolophus) in addition to those mentioned above. The Adélie penguin (Pygo- scelis adeliae) probably occurs, but was not recognized with certainty. Mecking's statement that the Emperor penguin (Aptenodytes forsteri) occurs at the islands in large numbers is certainly incorrect. The Cape pigeon (Daption capensis), known to some of the early voyagers as “‘the Spotted Eaglet,’’ is extremely abundant at most of the islands, but seemed to be less common at the Southern Thule Group. Very frequently it was seen flying or swimming with the Silver-grey fulmar (Prio- cella glacialoides) (= P. antarctica}, which is equally numerous. Both species were seen in thousands perched on ledges on the steep cliff sides, and we think it certain that they nest on the islands. Another species common at many if not all the islands is a tern, doubtless a form of Sterna vittata; small flocks were frequently seen sitting head to wind on an iceberg in company with Antarctic penguins. The Giant petrel (Macronectes giganteus) was sometimes numerous, but the Shoemaker (Majaqueus aequinoctialis), a skua which appeared to be identical with that at South Georgia (Catharacta linnbergi clarked) and the Dominican gull (Larus dominicanus) were less common. Wilson's petrel (Oceanites oceanicus), a most abundant bird at South Georgia, was here very scarce, and whale birds (Prion sp.) [= probably Pachyprila desolata] were seen only on a few occasions. A Shag (Phalacrocorax atriceps) [subspecies doubtful] was observed at the Candlemas Group, and a large flock of them was constantly around us while at anchor at Bristol Island. A single Mollymauk (Diomedea melanophrys) was seen at Saunders. Other Albatross (Diomedea exulans and Phoebetria), the Snow petrel (Pagodroma nivea), the Diving petrel (Pelecanoides) and the Sheath- bill (Chionis alba) were not seen at all. In the nesting season other birds are doubtless to be found at the islands: Larsen reports that in November he saw thousands of Pagodroma nivea nesting with Daption on Leskov (Kemp and Nelson, 1931, 156). c. South Orkneys. The South Orkneys lie about 500 kilometers from the tip of the Antarctic Archipelago, nearly 1000 kilometers southeast by east of the Falklands, and about half that southwest of South Georgia. Although landed upon frequer tly between 1821 and 1893, there were no more than meagre descriptions until the visit and subsequent wintering of the ‘Scotia’ in 1903 and 1904. Only two of the several islands are of considerable size, namely, Coronation, which is the most westerly, and Laurie, the most easterly. These and the lesser islets repre- sent parts of mountain ranges which have been submerged to such an extent that only their topmost peaks project from the sea. Lying on the northerly side of the antarctic low-pressure trough, the South Orkneys are constantly swept by west and northwest winds. During four- 226 OCEANIC BIRDS OF SOUTH AMERICA ay, SOUTH ORKNEY ISLANDS Miles 10 20 30 Kilometers 0 10 20 30 «440 50 60 Heights in meters : ‘ Ww, Inaccessible 8 ee : mS Hine I. ah eo ae Fic. 33. fifths of the year the sky is completely overcast; in this respect the group con- stitutes a sort of antarctic equivalent of Jan Mayen. Precipitation averages little more than one-third that of South Georgia. The mean annual temperature is about —5.0° C., but it varies greatly both diurnally and annually with the breaking through of polar influences. The mean temperature for the three warm- est months barely passes the freezing point (0.1° C. for January). Therefore the cold desert and tundra of the true antarctic reign here, limiting plant life to the development of mosses and lichens (Mecking, 1928, 318). Clarke (1913, 220) summarizes the relation of South Orkney weather to surrounding meteorological influences as follows: . . . in winter, owing to the freezing up of the sea to the south, the islands are virtually on the edge of a continent, and the temperature at that season is thus characterised by great varia- bility, the range of the thermometer frequently exceeding 60° F. in twenty-four hours. If the wind is in the south, very low temperatures, as low as 40° below zero F., are recorded; but with a change of wind to the north the thermometer may rise, even in the depth of winter, above the freezing- point. Summer is characterised by almost continuously overcast skies, and the finest and clearest weather occurs in winter. Owing to the large amount of cloud which hangs over the islands in summer, the temperature is much the same as at places ten degrees further south. The snowfall is excessive, the sunshine is very deficient, and strong gales are frequent. Laurie Island, where the ‘Scotia’ wintered, was the main scene of the labors of the Scottish expedition, and it is upon observations and collections made during eleven months’ residence there that the following notes are based. The length of this island is about 19 kilometers, its maximum breadth 10 kilometers, and its area fully 75 square kilometers. The interior is lofty, and several of the summits reach to an altitude of from 600 to 900 meters. A number of deep bays run inland from north to south, separated by narrow rocky peninsulas or steep ridges, and cause the island to have a very polypode outline. All the valleys EWE SCOT TAS ARG 227 are choked by glaciers, and what little exposed rock remains visible is precipi- tous in the extreme. Here and there on the lower slopes and at sea level are a few areas of more or less level ground. In winter the whole island, including even the faces of cliffs, becomes covered with snow, which does not commence to disappear until October and November (the late spring and early summer months); then, however, patches of moss-covered ground are laid bare, some of them supporting a thin soil of vegetable mould. The rocks are mostly covered with lichens, especially Usnea (Clarke, 1913, 219). The flora of the islands is poor in species, poorer than might be expected from their latitude. The predominant forms are lichens and mosses, but there are fewer species of both than in Graham Land farther west and south: fourteen mosses and eleven lichens are known from the South Orkneys, whereas twenty-seven mosses and fifty-five lichens have been found on the west coast of Graham Land, as well as two flowering plants and three hepatics unknown from the South Orkneys. There are, nevertheless, pleasing patches of vegetation in the lower snow-free places. Seals and sea birds are summer visitors that leave when the islands are ice-bound in winter. Fur-seals once occurred in small numbers, but none remains; there are Sea Elephants, Weddell Seals, Sea Leopards, and occasional Crab-eaters. Eighteen species of birds nest in the group, some in enormous numbers; there are Petrels, Gulls, Terns, a Shag, the Sheathbill, and Penguins. Adélie and Antarctic Penguins are very numerous. There are few invertebrates: two earthworms, some free-living Nematodes, three Collembolans, and a few mites are known (John, 1934, 30). Knowledge of the bird life of the South Orkneys is due mainly to the investi- gations of the Scottish naturalists, as reported by Clarke. However, Valette (1906, 3), who resided for a considerable period at the meteorological station maintained by the Republic of Argentina, has also given us an extraordinarily illuminating record of the life histories of South Orkney birds, a document upon which I have drawn heavily in preparing the accounts of several species. The birds known to breed at the islands are the following: PENGUINS Pygoscelis papua Pygoscelis antarctica Pygoscelis adeliae Eudyptes chrysolophus PETRELS CorMORANTS Macronectes Siganteus. Phalacrocorax atriceps Daption capensis So cua Ge Pachyptila desolata Rei ; : Chionis alba Priocella antarctica Thalassoica antarctica (?) Sxuas, GuLLs AND TERNS Pagodroma nivea Catharacta skua lonnbergi Oceanites oceanicus Larus dominicanus Fregetta tropica Sterna vittata The bird last named is listed in most of the literature as the South American Tern (Sterna hirundinacea). Elsewhere, however, I present evidence to show that the latter species does not enter the antarctic regions. 228 OCEANIC BIRDS OF SOUTH AMERICA SOUTH SHETLAND ISLANDS i Miles Oo 10 20 40 60 80 100 Kilometers eae 01020 40 60 80 100 Heights in meters Aspland |l. OBrien iF SNe I. . Bransfield Rks Foreland Dec ception |. a ar Sail! RS" ort a2 ter net d. South Shetlands. Just off the northern tip of the Antarctic Archipelago, which was until recently regarded as a mainland peninsula, are the South Shetlands which, together with Elephant and Clarence Islands to eastward, constitute a chain more than 500 kilometers in length. The group as a whole lies between the parallels of 61° and 63° S. The islands are wild and mountainous, with sharp peaks surrounded or covered by ice, and the rock is at least in part andesitic lavas which have counterparts in the backbone of South America. There is less floating ice, and far fewer bergs, than in the neighborhood of the South Orkney and South Sandwich groups because of the great stretch of open Pacific to windward of the South Shetlands, and the geographic and meteorological barriers between them and the waters of Weddell Sea. Deception Island, which is one of the few volcanoes that have been breached and inundated by the ocean, is the best known of the South Shetlands. The sheltered basin of its circular crater-harbor, with a single narrow entrance and a depth of 175 meters, was used by Yankee sealers of the early nineteenth century and is now the seat of an extensive whaling exploitation. The ice-covering of Deception Island is more or less undermined and elim- inated by volcanic gases, steam, and hot springs. Mosses and lichens form in some places a carpet over the soil, where insects of a few forms thrive. Grass is lacking. There are penguin rookeries with a total population of more than one hundred thousand birds. These, and other species, seem to be affected to a certain extent by the volcanic warmth of the island. Snow-free sites appear relatively early in the antarctic spring, and the Kelp Gull nests at Deception Island, according to Bennett (1927, 259), fully a month earlier than it does at the Falklands. THE-SCOTra( AEC 229 The composition of South Shetland bird life is not very different from that at the South Orkneys, but excessive numbers have been attracted to the former group by modern whaling operations. The still harbor of Deception Island is sometimes almost hidden under a raft of petrels of several sorts. Sir Hubert Wilkins, who made the island the base of his notable exploratory flights toward the antarctic continent, has informed me that the enormous bird population constitutes the greatest single hazard for an airplane. When taking off from the water, or alighting, at Deception Island, he has found it necessary first to attempt to clear a lane of water by driving a power-boat back and forth through the hordes of swimming birds. The fault basin of Bransfield Strait, which separates the South Shetlands from the neighboring antarctic region of many names, which is now known to consist of lofty islands, forms the pathway of numerous and variable ocean currents. One movement of surface waters, swinging around the tip of the Antarctic Archipelago, tends to thrust the ice-fields far back toward the south- west. A warm eddy from the north frequently passes southward through the strait, while close along the southerly coasts of the South Shetlands there is commonly a cold current running strongly northeastward. Here in late January the ‘Meteor’ found air temperatures of 1° C. at the surface of the water, overlaid by temperatures up to 11° C. at altitudes of 300 meters, a condition which sufficiently accounts for the dense, low-lying fogs that often prevail under clear skies (Spiess, 1928, 189). e. The Antarctic Archipelago. ‘Palmer Land,’’ ‘‘Graham Land,”’ and other units of this long projection of West Antarctica, succeed one another southward in the form of islands large and small to a point beyond the 70th parallel, where Stefansson Strait divides the chain from Hearst Land in the still problematical part of the antarctic continent. Through this gap it is probable that quantities of ice are short- circuited out of the head of Weddell Sea. The insular wall of the Antarctic Archipelago is steep along the western side, with a tabular region to eastward. Like Chilean Patagonia it has a fiord-and-channel coast. Old beaches, at both lower and higher levels than those of the present day, are perhaps as familiar here as in southern South America (Daly, 1920, 246). A mountainous country, with peaks at least 3000 meters in height, with snowfall greater than snow-waste, even during the summer months, and with the rock exposed only where the slopes are steeper than the angle of repose for snow, it is in some respects a region of extreme polar conditions. Terrestrial vegetation is reduced to a minimum, and away from the penguin rookeries there is almost no terrestrial fauna. Taking advantage of a milieu produced by the birds, there are mites, a few such lowly organized insects as Collembola, a wingless Chironomid fly, a number of rotifers, one or more tardigrades, and protozoans, in the moss. Members of several of these groups are, like the vege- tation, characterized by enormously specialized viability. They awaken for at most a few days in the summer season in order to carry on their active life 230 OCEANIC BIRDS OF SOUTH AMERICA nd Ss x Elephant}, ANTARCTIC ARCHIPELAGO Miles sO bo 150 720 sion ete = : aa 6 aes S Heights in meters Livingston | oe Srjow I. Smith |. * | Deception L ast — se “nao ase : °° Ro: we Antwerp | Ac a 7 Sey aisle Ow Hill |, Wiencke! Wandel I= me ts acy, Grob rt: (Mor pre hoe an L es Robingon . he ee Mobiloil Bay : eet Boast ‘ck Bowman Co, a Scripps I. HEARST LAND 68 66 és be Fic. 35. processes, and they are capable of existing for months or even for years in a frozen state. Despite this picture, the temperatures along the whole western coast of the Antarctic Archipelago are relatively high for the respective latitudes because of the tempering influence of the open sea to westward. The minimum tempera- tures, in particular, are less severe than in Weddell Sea or even at the South Orkneys. In more or less direct relation with these facts, it is noteworthy that the penguins and other antarctic birds all breed considerably farther southward along the west coast of the Antarctic Archipelago than they do on the eastern or Weddell Sea side. This is clearly shown by the distributional charts of the Second French Antarctic Expedition (Gain, 1914). THE SGOTIAGAREG 231 The present paucity of vegetation and animal life in this region is doubtless post-glacial, and is due not alone to the direct effect of cold but also to the physiological dryness of land covered with unusable water, 7. e. snow and ice. At any rate, a rich flora which resembled in many respects that of modern New Zealand, Australia, and southern South America, once flourished. Fossil floras of both Jurassic and late Cretaceous or early Tertiary age, and including such trees as Sequoia, Araucaria, and beeches, are known from West Antarctica. It has been suggested that from this region as a center many plants and animals once extended theirranges into more northerly zones. At Seymour and Snow Hill Islands, just east of the tip of the Antarctic Archipelago, the Swedish expedition discovered the extraordinarily varied fossils of middle Tertiary penguins which are briefly described below in my introduction to the section on that family. For much of our still scanty knowledge of contemporary bird life in West Antarctica we are indebted to the researches of the Second French Antarctic Expedition, commanded by Dr. Jean Charcot. By the members of his corps the neck of ‘‘Graham Land”’ and the adjacent parts of this most northerly extension of the south polar regions were not, of course, recognized as wholly insular. Nevertheless, between the years 1908 and 1910 the ‘Pourquoi Pas?’ cruised southward from Beagle Channel to Deception Island, turned northeastward to the end of Bransfield Strait, and subsequently skirted the entire western coast of the Antarctic Archipelago southward beyond the 70th parallel. The French ship next proceeded westward past lonely Peter Island, which is long and narrow, with a bare cliff suitable for the nesting of sea birds on its western face, and then on beyond longitude 120° W., before turning again toward the Magellanic district. According to the masterly report of Gain, ornithologist of the French expe- dition, only about 25 kinds of birds were observed beyond latitude 60° S., and most of these were wanderers from more northerly belts rather than polar species. As pointed out in several parts of my text, the antarctic Procellarii- formes have been quick to take’advantage of the new butchery produced by the whale fisheries. They assemble in vast numbers wherever shore whaling is in progress, and even appear to wait about for the steamers to arrive at their customary berths. Bennett (1927, 259), who has especially studied this phenom- enon, states that the incredible swarms of Cape Pigeons and other petrels have increased enormously since the year 1915, and that the area of their nesting sites has been proportionately enlarged. He believes, moreover, that the alba- trosses nesting at South Georgia and elsewhere in American Antarctica have become more local in their ranges since the advent of whaling, and that only two birds, the Snow Petrel (Pagodroma) and the Antarctic Petrel (Thalassoica), have been affected in no way by the novel conditions suddenly produced by man. With a handful of exceptions, all of the antarctic birds subsist, of course, only upon food which they themselves take from the water. The only species which find any portion of their food upon the land or which, indeed, are even capable of recognizing food that is not afloat, are the skuas, the Tern, the Gull, the Sheath-bill, and the Giant Fulmar. Penguins and petrels alike would liter- ally starve to death in the midst of plenty if their normal food of fish, or shrimps, 252 OCEANIC BIRDS OF SOUTH AMERICA or cephalopods, were heaped up about their nests and the birds restrained in some way from entering the water. A strictly new terrestrial feeding technique of Wilson’s Petrel (Oceanites) is, however, reported by Bennett, who states that these Mother Carey's chickens have acquired the custom of feeding upon coagu- lated whale-oil that washes ashore. The birds have therefore learned the new art of ‘‘walking’’ on the land while feeding. They accomplish this by keeping their wings vibrating in exactly the same manner as though they were skipping and dancing across the surface of the water. In considering the following list of birds that cross the 60th parallel of south latitude, a list modified from the records of the ‘Pourquoi Pas?’ naturalists, the place of whaling operations as a somewhat artificial lure must be kept in mind. Forms known to breed in the Antarctic Archipelago are preceded by an asterisk. The exact identity of a number of subspecies, and even of species, is still open to doubt. Such petrels as Procellaria, which has been observed south- ward to De Gerlache Strait and over the open Pacific to latitude 63° S., and Adamastor, recorded as far as latitude 60° S., are mere wanderers. PENGUINS Aptenodytes forsteri (migrates northward along both coasts of the Antarctic Archipelago, and doubtless breeds in high latitudes both to eastward and westward) Aptenodytes patagonica (probably bred in former times as far southward as the South Shetlands; see page 344) *Pygoscelis papua (breeds southward to about latitude 65° 30’ S.) *Pygoscelis adeliae *Pygoscelis antarctica (breeds southward to about latitude 65° S.) ALBATROSSES AND PETRELS Diomedea exulans Diomedea melanophris Diomedea chrysostoma Phoebetria palpebrata *Macronectes giganteus (breeds southward to De Gerlache Strait, about latitude 64° 30’ S.) *Daption capensis Halobaena caerulea Pachyptila desolata *Priocella antarctica *Thalassoica antarctica Adamastor cinereus Procellaria aequinoctialis *Pagodroma nivea *Oceanites oceanicus CorMORANTS *Phalacrocorax triceps (breeds southward slightly beyond latitude 65° S.) SHEATH-BILLS *Chionis alba Skuas *Catharacta skua linnbergi *Catharacta skua maccormicki (the dividing line between the breeding ranges of the two forms of skua is still prob- lematical) GuLLs *T arus dominicanus TERNS *Sterna vittata THETSCOPTAS ARC 233 The identity of the last species, the Antarctic Tern, has been confused in much of the ornithological literature pertaining to the region. In Gain's (1914, 184) report, however, the species is correctly named and is recorded as a breed- ing bird at the South Orkneys, South Shetlands, Louis Philippe Land, Seymour, Snow Hill, Joinville, and Paulet Islands, ‘‘Graham Land,’’ Booth-Wandel Island, and other antarctic localities. Gain states that the Polar Skua (Catha- racta skua maccormicki) also nests northward to the South Shetlands, where it meets and mingles with the lower antarctic form (C. s. lénnbergé). This problem is discussed in the biographies. It may yet prove necessary to regard the two skuas as distinct species, instead of as races of a common species. (1) BOUVET ISLAND Bouvet Island is far from the South American region but it is important from our point of view as a way station in the eastward distribution of two or more species of sea birds which probably originated in the American quad- rant of Antarctica. The island was one of the first antarctic land masses known to Europeans, being discoy- ered in 1739 by Pierre Bouvet, who mistook it for a headland of a southern continent. Its subsequent history was for a long time puz- : zling. Often searched for in vain, | 4°992rds rediscovered and lost again, it ap- peared upon successive maps not only in various positions, but even as several distinct islands, until its latitude and longitude were cor- rectly ascertained in 1898 during the cruise of the ‘Valdivia.’ The island rises abruptly from great depths to an elevation of 935 meters above sea level, and it is almost completely covered by an ice cap with sheared and precipitous edges around most of the coast. On the western side are steep cliffs with an altitude of about 400 meters. Although the latitude of the island is the same as that of South Georgia, its glaciation is considerably more excessive because of the influence of Weddell Sea ice which here follows the general trend of wind and current circulation toward the north and east. The ‘Meteor,’ on her transoceanic cruise from the Falklands, encountered near the mid-Atlantic rise this drift of very cold water, which sharply lowered the temperature of the air (Spiess, 1928, 138). Bouvet is surrounded by pack-ice during much of the year and is prevailingly hidden by dense fog. Dredgings from the adjacent ocean bottom, and observations ashore as well as from an airplane, BOUVET ISLAND fo) 1 2 3 4 5 Miles o t 2 3 4 _ 5 Kilometers Heights in meters C.Valdivia C. Lollo 935 Kaiser Wilhelm Pk rise “ena CL rsl. Sens Gy Middle of Island, Lat. 54°26 S., Long 3°24'E. Fic. 36. 234 OCEANIC BIRDS OF SOUTH AMERICA show that the island is volcanic. Sachse, of the ‘Valdivia,’ presumed it to be a cone, with traces of a crater near its high northern side, but Mosby, of the ‘Norvegia’ expedition, regarded it as a plateau, the whole mass, according to his view, being part of one flat stream of lava. The visible discharge of gases from Bouvet, reported by the observers of the ‘Meteor’ in 1926, was not ob- served by the ‘Norvegia’ party (Aagaard, 1929, 71). Tosouthward and eastward are relatively gentle slopes toward the level of the sea. The east-west diameter of the island is about 9.5 kilometers. On December 1, 1927, when the ‘Norvegia’ expedition took possession of Bouvet in the name of the King of Norway, countless penguins of “‘three species’ were observed at the commencement of their breeding season (Aagaard, 1928, 5). Photographs have identified two of these as the Ringed Penguin (Pygoscelis antarctica) and the Macaroni Penguin (Exzdyptes chrysolophus). The third has been called the Adélie Penguin but, in the absence of proof, the presence of this high antarctic species at Bouvet must be regarded as doubt- ful. If the Adélie does prove to be a resident, the record will mark a notable northward extension of the range, and will emphasize the peculiar hydro- graphic and climatic conditions described above. Noteworthy, in the mean- time, is the fact that at Bouvet the Ringed Penguin, indubitably a bird of the American section of the Antarctic, here reaches the easternmost limit of its range, while the Macaroni Penguin has probably followed the same route in its partly circumpolar dispersal. By the Norwegian visitors penguin colonies were found both on the main island and on Lars Islet off the southwestern tip. The latter has a length of about 500 meters, is fairly flat and low, and has several tunnels completely through it. It is the home of great numbers of fur seals as well as of the penguins. Chun (1900, 171) records numbers of Cape Pigeons (Daption) and Snow Petrels (Pagodroma) in the vicinity of Bouvet, and the ‘Meteor’ party observed Wandering Albatrosses to southward of the island (Hentschel, 1933, 121). Data on the breeding birds are much needed. Reichenow (1904, 347), as a result of the ‘Valdivia’ collections, records Phoebetria palpebrata, Priocella antarc- tica, and a skua as resident birds. It would be of special interest to learn which of the following genera, and what forms of each, are represented at Bouvet: Pachyptila, Fregetta, Pelecanoides, Phalacrocorax, Chionis, Sterna. . EasteRN Furecia TO CENTRAL CHILE a. Staten Island. 9 Divided from Tierra del Fuego by the Strait of Le Maire is Staten Island, with its high ridge extending from west to east along the path of the Scotia Arc. In the words of de Agostini (1924, 228), the isle is a chaos of rugged rocks and peaks, almost without level terrain and with scarcely a shelving strand upon which a boat might safely run ashore. Rich beech woods and shrubbery of Fuegian type spread a green blanket over the mountain slopes. The island is furrowed—even almost cut through—by many narrow fiords, of which the most important is Puerto Parry, near the middle of the northern shore, and TIERRA DEL FUEGO AND SOUTHERN CHILE 235 easily recognized because Monte Buckland (914 meters) rises from its eastern side. Staten Island is a focus of such raging, buffeting winds as are to be expected in latitude 54° S. of the Atlantic. Its mean annual temperature is 5° C., which is very markedly higher than that of South Georgia, on the same parallel, but the seasonal succession of temperatures and rainfall is similar. An annual pre- cipitation of 610 millimeters is vastly less than that of the western or windward face of Tierra del Fuego. On the other hand, it is practically double that of Magallanes, showing how even a narrow reach of ocean, coupled with an abrupt wall of hills, can make possible the wringing of new moisture out of winds so recently depleted. Not far off Puerto Parry are the five small Afio Nuevo Islets, on the largest of which meteorological observations have been recorded for many years. Here Argentine mail ships call regularly. Large colonies of ocean birds occupy Afio Nuevo as well as the northeastern end of the main island, where a peninsula known as Punta Pajaro (Bird Point) extends from a treeless stretch of peatland. This is occupied by thousands of penguins, cormorants, and other Magellanic species, while sea-lions and fur seals haul out on the rocks at sea level. The avifauna of Staten Island cannot be called well known, but it doubtless includes most of the sub-antarctic sea fowl of Tierra del Fuego and the Falklands, as well as numerous land birds. Four kinds of penguins have been recorded as resident, together with albatrosses, petrels, shags, skuas, gulls, terns, shore birds, and ducks. In addition to the casual notes of de Agostini, data on the birds have been published by Salvadori (1900), Crawshay (1907), and Hartert and Venturi (1919). In September, 1915, Mr. Beck collected or observed in the waters about Staten Island albatrosses of four species, petrels of the genera Macronectes, Daption, Priocella, Halobaena, Procellaria, Pachyptila, and Garrodia, several of which were merely transients, as well as two species of shags and the Sooty Shearwater (Puffinus griseus). The last presumably nests at the island. b. Tierra del Fuego and Patagonian Chile. Quien come Califate, Vuelve por mas. Of one of the earth’s most complicated minglings of land and sea, in which the islets are numbered literally by thousands, I can perforce give only a cursory description. For a suggestive cross-section of the region, as made during the Brewster-Sanford Expedition, the reader is referred again to the summary of Mr. Beck's narrative, pages 8 to 25 of this book. In the eastern part of the area is the great land mass of Tierra del Fuego itself, the leeward portions largely low-lying and relatively dry. To southward it is nearly cleft in twain by the long arm of Admiralty Sound and the higher plane of Lake Cami. Beyond these, in turn, stretches the lofty Cordillera Darwin, which terminates to westward and windward in the jig-saw puzzle of the Brecknock Peninsula, where mainland and islet can scarcely be distin- “LE “OL Zauiwey o8aig a : a oy] | svagaw ur sayBioy | SONYOH 3a Ogvo — 9S Cpl eS Hens aaes ae f SuaqPaWo}iy SOUJOH IN, aqilual 001 L os sz Ja0aq DR es nal P-OSUOJAP|| Sy Saji bea fate we SOUJOH ap oge> osjey : Serif SIapUay + mee we OY “PBuoqsy g O94N4 130 vayaiL Apsey - Noss. f jel] ae) r 4 ey) ‘ Q -Sowysinyy ouag erany'1/ ) / Pie = 4 oy | ay Oulueneny +} Ae OF i ASF NAA aydrogh 6 10009 SM aA TUR hOqUemays “| QS Cyuopuoy | a N UINGYI0) /0URD onany OUuY s} S| ylW3H9 8 ty | 1S] Pp. s oe [ 236] =e ol Uo!dejOSaq “| 9%, eae LX OuBWEdS W 5 Gf of are a "Yasgo7 aiPps Pt [0 oy cid wl al UNAS OUaS "uod = TPES ceiqewaseny odnug ae oF +Vcaaured ‘1 | ces Ae ENN, 2 | a i A “Ete a ou LN Zl LA) his Soe SS etate 9b TIERRA DEL FUEGO AND SOUTHERN CHILE 237 guished one from the other. The same is true, indeed, of the labyrinths still farther southward, across Beagle Channel, and of those that extend northwest- ward to intercalate with ameboid branches of the drowned end of a continent. Let us picture the region briefly in the words of those who have traced its tangles and felt its spell. Of all the regions of the earth's temperate zones it is doubtful whether there remains a single one of which man's knowledge is so incomplete as of that of the southern Patagonian Andes from the forty-fifth parallel of latitude to the outermost islands of the Tierra del Fuego archipelago. Yet we can hardly call this land of mystery remote: not one of its 180,000 odd square miles of mainland and islands is 200 miles distant from one of the New World's oldest trade routes. Nor has there been a dearth of interest in it. Between Magellan's discovery of the strait in 1520 and the visit of the Beagle in 1832 no fewer than 81 expeditions representing many nations have ex- plored there. Between the voyage of the Beagle and the publication of ‘‘Charles Darwin's Diary of the Voyage of H. M.S. ‘Beagle’ ”’ last year scores of books and articles in various languages have ap- peared. Yet one needs only to glance over the most recent literature to realize how little is actually known, and even the latest maps disclose vast areas marked ‘‘inexplorado’’ (Rudolph, 1934, 251). The external features of Tierra del Fuego are exceedingly varied. There are lowland flats with vast marshes and lakes more or less brackish, scrub-covered downs, bleak black peaty moors, practically impenetrable forests, and regions of everlasting snow probably never trodden by man. In the coastline, there is also much diversity. This may be low shingly beach with the land dead flat behind it, or bare perpendicular cliff washed at foot by the sea, or solid jagged rock overgrown to high-tide mark with impermeable thorn scrub, or else precipitous mountain covered with forest or glacier to the very ocean. Grey sky, grey sea, grey beach, the ]and white, and the black rocky crests of snow mountains standing out in threatening relief suspended between earth and heaven. What a study the seashore, here, where the two greatest oceans meet and sweep round the tail of the greatest continent! What tremendous force of wind and water! Piled up in such confusion as to make one stand aghast in contemplating it are masses of sea-weed with rocks attached, mussel and limpet shells, the bones or entire carcases of whales—large and small, the carcases of sea lions and of guanacos, trunks of trees, and such evidence of the existence of man as a boat, a spar, an oar, a companion ladder, a ship's draught-board painted in black and white chequers on a stout piece of plank, and other wreckage (Crawshay, 1907, xiii, vii). Among the earth's most turbulent waters are the seas of this region. The complicated currents set up in these labyrinths of canals by the ebb and flow of tides having considerable range were compared by Darwin to movements in a tub of boiling water, and the remains of vessels found along these rocky channels tell their own story. Diurnal tidal ranges here are equaled at but few coasts that at Puerto Gallegos, 45.6 feet (nearly 15 meters), being the greatest of all the world’s tides listed by Bauer . . . . The height of the waves in these regions is also noteworthy, reach- ing 15 meters and more at Cape Horn. Employees of the lighthouse at one of the Islet Evangelistas, at the western entrance to the strait, found it necessary to build rampart walls to protect their water supply, at 56 meters height, from salt-water spray. The west of the cordillera ... is . . . one of the wettest regions of the earth's surface. Rainfall is nowhere less than 1000 millimeters; it reaches over 5000 millimeters ... on Isla Desolacion. ... In the western and southern regions rain is more or less continuous throughout the year. Admiral] Spiess writes of the Cockburn Canal that it has but a single day in a hundred of calm, sunshine and only occasional precipitation, while of Staten Island it has been written that 24 continuous hours without rain, sleet, or snow are unknown... . . Patagonia is also known for its “trenchant wind.’’ The tip of the continent is marked by a conflict of ocean currents, the cold westerly drift dividing at Cape Horn into northward and east- ward-flowing branches, while a branch of the warm current from Brazil impinges seasonally upon the Atlantic littoral. Turbulent wind conditions result—not alone the normal high winds due to an abnormally low average barometer but storms that arise suddenly and rage for a week or an entire month (Rudolph, 1934, 260). 238 OCEANIC BIRDS OF SOUTH AMERICA Only when the winds have a southerly slant does bright, clear weather reign. Sometimes on quiet days banks of white, sharp-edged, solid-looking cumulus clouds may be seen mounting in the south, indicating to mariners a rising of the southwest wind. The clearest months in the Strait of Magellan are from September to the end of December, and the worst, on account of snow and heavy gales, from the middle of June to the middle of August. Because of sketchy charting, tortuous bends, and innumerable reefs, naviga- tion at night is rarely attempted in the Fuegian channels. Fortunately, friendly coves and holding ground are plentiful. Snug in such shelter, many voyagers have heard the storms raging fiercely overhead, while only occasional and fitful gusts descended as low as their ships. Thus, lying at anchor, Cordova (1820, 21) could write, ““All the night the weather was cloudy, with showers of rain or rather snow. The wind continued with most furious blasts, which, however, only affected our upper rigging, without reaching the hull of the vessel.”’ [There is a sharp line of division] between the plant zones in southern Patagonia. The border between pampa vegetation and the forest approximates the eastern limit of the cordilleran-forma- tions. . . . The plains of eastern Patagonia and of northern Tierra del Fuego are covered with short bunch grass and xerophytic bush. To the west lie the forest zone of deciduous beeches (Nothofagus antarctica on the lower slopes and Nothofagus pumilio on the slopes above 300 meters) and, finally, the moist zone of evergreen beeches (Nothofagus betuloides). Between is a narrow transition belt of underbrush, where the pampa invades the forest at one point and vice versa at another (Rudolph, 1934, 261). Darwin speaks of the attractive, park-like appearance of the country in southern Tierra del Fuego, where the grasslands begin to give way to forests. Below is his field notebook impression of heavily wooded country on Brunswick Peninsula, about midway through the Strait of Magellan: I left the ship at four o'clock in the morning to ascend Mount Tarn; this is the highest land in this neighbourhood being 2,600 feet above the sea. For the two first hours I never expected to reach the summit. It is necessary always to have recourse to the compass: it is barely possible to see the sky & every other landmark which might serve as a guide is totally shut out. In the deep ravines the death-like scene of desolation exceeds all description. It was blowing a gale of wind, but not a breath stirred the leaves of the highest trees; everything was dripping with water; even the very Fungi could not flourish. In the bottom of the valleys it is impossible to travel, they are barricaded & crossed in every direction by great mouldering trunks: when using one of these as a bridge your course will often be arrested by sinking fairly up to the knee in the rotten wood; in the same manner it is startling to rest against a thick tree & find a mass of decayed matter is ready to fall with the slightest blow. I at last found myself amongst the stunted trees & soon reached the bare ridge which conducted me to the summit. Here was a true Tierra del Fuego view; irregular chains of hills, mottled with patches of snow; deep yellowish-green valleys, & arms of the sea running in all directions; the atmosphere was not however clear, & indeed the strong wind was so piercingly cold, that it would prevent much enjoyment under any circumstances. I had the good luck to find some shells in the rocks near the summit. Our return was much easier as the weight of the body will force a passage through the underwood; & all the slips & falls are in the right direction (Darwin, 1933, 208). Lord Bryce (1912, 306) writes: Magellan's Straits are unlike any other straits in this respect, that the physical aspect of the two ends is entirely different. The character of the shores on each side is the same in each part TIERRA DEL FUEGO AND SOUTHERN CHILE A) of the channel, but both shores of the eastern half, from the Atlantic to Cape Froward, are unlike those of the western half from Cape Froward to the Pacific. The former has low banks, with smooth outlines, slopes of earth or sand dipping into shallow water, and a climate extremely dry. The latter half is enclosed between high, steep mountains which are drenched by incessant rains. The eastern half is a channel, narrow at two points only, leading through the southernmost part of the vast Argentine plain, which has apparently been raised from the sea bottom in comparatively recent times. The western half is a deep narrow cut through the extremity of a great mountain system that stretches north for thousands of miles, forming the western edge of South America, and the rocks on each side of it are ancient (palaeozoic or earlier). The western half is grand and solemn, with its deep waters mirroring white crags and blue glaciers. The low eastern half has no beauty save that which belongs to vast open spaces of level land and smooth water over which broods the silence of a clear and lucent air. A more singular contrast, all within a few hours’ steaming, it would be hard to find. Cape Froward, at the bend of the Strait of Magellan, marks the point in this latitude at which the mountains give way to the plains. East of this point, continues Bryce, who had come from the west, . One is at once in a different region with a different climate. The air is drier and clearer. The shores are lower, the wood, still mostly of the Antarctic beech, is thicker, with many dead white trunks which take fire easily. The hills recede from the sea, and grow smoother in outline, finally disposing themselves in low flat-topped ridges, six or eight miles behind the shore-line. A wide expanse of water, and of land almost as level as the water, stretches out to the eastern horizon, so that at first one fancies that this apparently shoreless sea is part of the Atlantic, which is in fact still nearly a hundred miles away. [Magallanes is the same distance to the south of the equator as the Strait of Belle Isle in Labra- dor is to the north], but the climate here is far more equable. It is never warm, but the winters are not severe, there is little snow, and frosts are moderated by the adjoining sea. The air is dry and healthy with a rain-fall of only ten inches a year. Though the landscape is bare, for trees can with difficulty be induced to grow, and though there is much wind and no shelter, still we found something attractive in this remote and singular spot, for one has a constantly stimulative sense of the vast expanse of sky and sea and the distant plain of Tierra del Fuego, with a touch of mystery in the still more distant ranges of that island which just shew their snowy peaks on the horizon. The light over sea and shore has an exquisite pearly clearness which reminds one of the similar light that floats over the lagoons between Venice and Aquileia. Can this peculiar quality of the atmosphere be due, here as there, to the presence of a large body of comparatively smooth and shallow water, mirroring back to heaven the light that it receives? Crawshay also dwells upon the subdued tone of everything to be seen in land and sea and sky in Tierra del Fuego, and writes that ‘‘the soft sunlight resembles bright moonlight rather than the light of day.” Climatic conditions in Beagle Channel more or less parallel those in the Strait of Magellan. Thus to eastward of Ushuaia the weather becomes milder and dryer than in the western reach. The soil is less marshy and bears a rich pasturage. Grandeur in the landscape is replaced by aspects which may rather be described as pleasant and cheerful, in part because man has here built roads through the forests and dwellings along the shores. Gable Island, near the eastern end of the channel, consists of fertile alluvial land that supports several thousands of sheep. Opposite is Puerto Harberton and the beginning of the pack trail to Lake Cami, where the breeding of Royal Albatrosses (Diomedea epomophora) remains to be confirmed. The world’s southernmost farm lands are on Navarino Island, just to southward. Though mountainous and heavily 240 OCEANIC BIRDS OF SOUTH AMERICA wooded, this island supports thriving sheep estancias along its northern coast. Navarino is the only island beyond Tierra del Fuego on which guanacos live, and wherever this mammal occurs life has proved tolerable for human beings, as regards both climate and productivity of soil (de Agostini, 1924, 218). Father de Agostini has described also the picturesque northeast arm of Beagle Channel, which lies between the Cordillera Darwin and Gordon Island. On both sides the mountain steeps are here smothered under a growth of beeches and Jaurels, all of a fresh and bright shade of green. Higher up are vast stretches of yellow and reddish cryptogams, lined by glittering brooks and white cas- cades that are swallowed from view when they reach the forest. Gleaming snow covers the peaks beyond. Numerous glaciers here descend to be reflected in the water of Beagle Channel, and small bergs against a background of greenery make a spectacle to astound the seafarer. Most majestic among these streams of ice is the Italia Glacier, with a menacing front more than a thousand meters in length along the channel, and a source on the slopes of Mount Darwin (2135 meters). Two other modern authorities have written as follows regarding Fuegian glaciation: The most striking feature of Beagle Channel is undoubtedly the astonishing number and variety of the glaciers, which occupy all the valleys descending from every mountain high enough to be covered with a mantle of perpetual snow. The general direction of these valleys is favourable as regards shelter from the summer sun. This fact, no doubt, accounts in a measure for the great extension of the glaciers. The coolness of the summer rather than the severity of the winter is also an important factor in maintaining both the glaciers and the comparatively low snow-line, which cannot be much more than 2000 feet above sea-level. It is to these features we must look in endeavouring to explain the glaciers of Tierra del Fuego. Most of the larger glaciers we saw showed signs of shrinkage. Trees grow-on the lower and older terminal moraines, the vegetation becoming younger as the actual ice-wall is approached. Lateral moraines are found stranded on the flanks of the valleys, high above the present level of the ice, and the polished rock indicates the height to which the glacier had once reached. These features were particularly noticeable in the case of the glacier opposite Romanche bay, and the large recession of glaciers in these regions is an interesting one for future investigation, and Beagle channel forms an admirable field for study (Crosthwait, 1905, 288). The permanent ice fields of southern Patagonia are the largest glacial areas found within the temperate zones. At their northern limit, near the Rio Aysen, these fields begin at about the 2000-meter level; in Tierra del Fuego they extend upward from 750-900 meters. Only a few crests and rocky points rise above these rugged white masses, which attain 30 kilometers and more in width. In the distance of 1200 kilometers from its beginning south of the Rio Aysen to its termina- tion with the Cordillera Darwin in Tierra del Fuego this ice mantle loses its continuity at but two localities—at Baker Canal and around the Strait of Magellan. Thus it is subdivided into three principal ice areas, of which the central is the most extensive. Ribbons of ice extend to the sea on the west and to the Andean lakes on the east. It is at Jatitude 46° 40’ S. that the northern- most of these glaciers, San Rafael, reaches the sea—10 full degrees more distant from the pole than the southernmost glacier of Alaska, more than 20 full degrees farther than Jokelfiord (68° N.), the southernmost glacier of Norway reaching the sea (Rudolph, 1934, 251). Except in the near vicinity of glacier-snouts, the temperature of the salt water in channels throughout the Magellanic region is said to be everywhere above 4.4° C., doubtless because of the effect of warm Pacific surface waters that work in with the westerly winds, which prevail throughout the year. TIERRA DEL FUEGO AND SOUTHERN CHILE 241 Reynolds (1935, 65, gives a charming account of a visit during December, 1932, to the islands of the Cape Horn neighborhood, which include Grévy, Baily, Wollaston, Freycinet, Hermite, Jerdan, Hall, Herschel, Deceit and Horn, together with numerous adjacent islets and the easterly outliers of Barnevelt and Evout. The main group makes a rather compact cluster to eastward of False Cape Horn on Hoste Island, and south of the Gulf of Nassau. Parts of these islands are comparatively flat, with peaty ground reminiscent of the Falklands. There is also, however, considerable high land, such as Mt. Hyde on Wollaston, which attains an altitude of 674 meters, as well as imposing sea cliffs with great pillars of basalt. The larger islands are, indeed, so high that patches of snow persist toward their summits throughout the summer. The Barnevelt Islets, 14 kilometers to eastward of Deceit, differ much from the others, consisting mainly of tussock-grown land completely occupied by bur- rowing Magellanic Penguins, and surrounded by rocks and crags which are covered with cormorants. All of these islands are now uninhabited, but are visited by otter hunters. The exposures toward the prevailing southwest wind on the Wollaston group are particularly bare and weather-beaten. Areas of outcropping granite pro- trude from a blanket of dwarfed and battered antarctic beach, through or over which progression is almost impossible. White-looking vertical lines of weather-killed trees are a feature of these gnarled woods. The north-eastern declivities are generally less steep, with the thick low vegetation covering peaty tracts, broken by bog-pools and extremely wet ground of a more open character. In sheltered places, such as the proximity of a lake at the northern side of Freycinet, precipitous ravines are submerged by evergreen forest of tropical aspect. So thick is the canopy that a chaos of decay exists in the semi-darkness prevailing beneath, and a small maidenhair fern combines with a deep layer of spongy moss to cover everything. In such an environment of wooded glens, Reynolds heard the strange, mim- icking notes of the Magellanic Thrush (Turdus magellanicus). Thus in the heavy drip of a rainy morning at Deceit Island, after the departure of petrels for sea, the subdued song of the thrush first unmistakably resembled the calling of these sea fowl as if at a distance, after which it imitated in turn the night voices of snipe and, at daybreak, its discreet whispering gave place to loud and bold mimicry of the chorus of Black Oyster-catchers, or the sudden clamor of Kelp Geese. Exclusive of more than 30 species of sea fowl, waders, ducks and geese, no less than 25 kinds of land birds inhabit these blustery islands. They include, in addition to the thrush, a house wren and a marsh wren, a swallow, a siskin and three other finches, an Icterid or American blackbird, three species of tyrant flycatchers, four of Cinclodes, a spinetail, a babbler (Scytalopus), a kingfisher, and six birds of prey, the latter including a buzzard, eagle, kite, two caracaras, and the Turkey Vulture. The shores of all of the islands are frequented by sea- lions, fur seals, and the Fuegian otter. The only land mammals are a bat and a rodent (Ctenomys?). The following two descriptive items relate, respectively, to the visit of the 242 OCEANIC BIRDS OF SOUTH AMERICA ‘Valhalla’ to the western end of the Strait of Magellan, and to the departure of the ‘Beagle’ from this exit into the Pacific: At daybreak . . . we left Port Gallant, and, proceeding westwards, anchored towards evening at Churruca Bay, which is the most western anchorage in the Straits of Magellan. This place is one of the most beautiful that it has ever been my good fortune to see. Towering peaks, covered from summit to base with impenetrable beech forests, almost encircled the deeply indented bay, the water of which was inky black, and the surface smooth as glass. Here and there were little islets, some fringed with a tall white-flowered plant; others surrounded with hedges of fuchsias in full flower. Flock after flock of Magellan cormorants flew by us, their white breasts flashing bright against the dark water of the bay, while steamer-ducks scudded to right and left, as we glided to our anchorage in the glorious natural harbour (Nicoll, 1908, 173). In the morning in company with the Adventure, we made the best of our way into the open ocean. The western coast generally consists of low, rounded, quite barren hills of Granite. Sir J. Narborough called one part of it South Desolation, ‘*because it is so desolate a land to behold”’; well indeed might he say so. Outside the main islands, there are numberless rocks & breakers, on which the long swell of the open Pacific incessantly rages. We passed out between the “East & West Furies’’; a little further to the north, the Captain from the number of breakers called the sea the ‘Milky Way.”’ The sight of such a coast is enough to make a landsman dream for a week about death, peril & shipwreck (Darwin, 1933, 230). Straight out to sea beyond the Long Reach are the rocks which Narborough called the Islets of Direction, but which are now known as the Evanjelistas. The group is conspicuous because of a hole through one of the slate cliffs and a contorted stratum of white quartz along the face of another. Rainfall at the lighthouse amounts to 2843 millimeters per annum. So uninterrupted are the hurtling, seething waves that, according to tradition, a vessel was once obliged to wait for forty days in the shelter of neighboring islets just off the coast of Pacheco Island before it could approach the Evanjelistas Rocks. Thus was established the name of the group and port still called Cuarenta Dias. According to Crawshay (1907) the majority of the birds in the Fuegian region are migratory summer visitors rather than permanent residents. This applies even to such common breeding waterfowl! as ducks and geese and, of course, to the kinds of petrels and their allies that nest among the islands. The orders represented in Crawshay’s book, as native to Tierra del Fuego, include penguins, grebes, cormorants, petrels and albatrosses, gulls, terns, and shore birds, rails, ducks and geese, herons, woodpeckers, parrots, owls, hawks, and many pas- serine forms. Chapman (1933, 371) thus describes a late January morning’s observa- tions, made in part from a motor car, in Tierra del Fuego, close to the Strait of Magellan: Our road, the product of use, crossed widespreading, rolling plains dotted with the yellow bloom of califat bushes and grazing sheep. Scattered here and there were families and flocks of Wild Geese. Eight Geese, it was said, eat as much grass as one sheep and these birds have no friends among raisers of wool and mutton . . . Strong-winged, heavy-bodied, powerful Slzuas' swept by us seeking what they might devour and Milvagos looked for lesser prey. There were Ducks and Gulls on the many ponds and lagoons, and shore-birds along their margins. . . . The sun rose on a brilliantly clear morning. I could plainly see details on the shores of the mainland. The air rang with the cries of Gulls, the whistling and cackling of Geese, the high TIERRA DEL FUEGO!) AND SOUTHERN CHILE 243 piping of Oyster catchers, the tinkling choruses of Ibises (Theristicus), the mellow, purring whistle of Widgeons, leading their broods on the lagoons, and from overhead, a Pipit showered me with song. There were notes in his voice that reminded me of both Skylark and Bobolink. [ lay on my back to watch him flutter upward and float downward. The channels of the southerly and windward coasts of Tierra del Fuego are frequently invaded during the winter by vast flights of oceanic birds, such as the Cape Pigeon, the Silver-gray Fulmar, and other wanderers from the Antarc- tic Zone. During the breeding season, however, one sees in these waterways few birds from offshore, but rather the native Kelp Goose and steamer ducks, cormorants, penguins, oyster-catchers, and skuas. After nightfall, new voices and the whish of wings may indicate the coming and going of the petrels that nest in this region, and about which next to nothing is yet known. Among the biographies forming Part II of this book I have quoted various obscure refer- ences to nesting petrels of Tierra del Fuego, such as Darwin's account of the whale-birds (Pachyptila) of Landfall Island, south of the western entrance to the Strait of Magellan. Beck's extraordinarily interesting discovery of the South American breeding grounds of the Sooty Shearwater and Wilson's Petrel is also recounted both in the life histories and in the narrative of the Brewster- Sanford Expedition. No doubt a rich field for ornithological finds, and data for solving many fascinating problems of avian taxonomy, distribution, biology, and behavior, awaits a field-worker prepared to investigate systematically the weather-beaten outer islands, such as Morton, Waterman, Gilbert, Stewart, Noir, Santa Inés, Nafios, Graves, Desolaci6n, and others, which stand as shock- troops against the assault of the gales from the Horn to Cape Pilar and beyond. The outer Patagonian coast, from the Strait of Magellan northward to the far end of the chain of islands formed by the submerged southern portion of the maritime cordillera, is made up of a succession of high, forbidding, steep- cliffed islets separated by innumerable channels and penetrated by countless fiords and inlets, with two great gulfs that break the chain altogether and bring the open sea to the mainland of the continent. By no means all of this compli- cated coast is yet fully surveyed, and information about certain ship-passages is based upon observations by Captain Stokes of the ‘Beagle’ and other explorers of the same remote period. A boat-journey through such routes as the Smyth, Sarmiento, Messier, and Darwin channels is, as Chapman says, ‘‘a voyage through the Andes.”’ The islands are so intricately cut up by reticulated water- ways that they are like the tiles of a mosaic. It is difficult to group them into archipelagoes and their very nomenclature has been subject to endless confusion. In general, the seaward faces of the outer islands have a very different appear- ance from their eastern sides, along the channels, the former tending to be bare and gray, the latter green and dark. Still farther eastward the wooded slopes appear almost black beneath the rocky crests and the ice caps of regions still partly ‘‘inesplorado.”’ To the traveller, steaming up the inner channel of this coast with its labyrinth of fjords and islands, the land appears hopeless for the uses of man. Glaciers crowd down the sides of precipitous mountains that are gloomy with almost perpetual mist and rain, and covered at every possible 244 OCEANIC BIRDS OF SOUTH AMERICA point with an evergreen forest. . . . The persistent west wind gives a heavy rainfall . . . . The explorer is sometimes delayed by two or even three weeks of steady rain, and progress in the forest must be made gver a mucky mass of rotting vegetation into which a man often sinks to his knees and where beasts of burden cannot go. Permanent settlements are not likely to be made here (Smith, 1904, 16). During the middle of the eighteenth century, Commodore Byron, grand- father of the poet, was wrecked in the ‘Wager’ on the coast of the Guayaneco Islands, which form the southerly boundary of the Gulf of Pefias. His de- scription of the region is as apropos today as it was at the date of Lord Anson's expedition round the world, in which the ‘Wager’ was participating when lost. The Country hereabouts wears the most uncouth, desolate, and rugged aspect imaginable; it is so circumstanced as to discourage the most sanguine adventurers from attempts to settle it; were it for no other reason than the constant heavy rains, or rather torrents, which pour down here, and the vast sea and surf which the prevailing westerly winds impel upon this coast, it must be rendered inhospitable. All entrance into the woods is not only extremely difficult, but hazard- ous; not from any assaults you are likely to meet with from wild beasts; for even these could hardly find convenient harbour here; but from the deep swamp, which is the reigning soil of this country, and in which the woods may be said rather to float than grow; so that, except upon a range of deformed broken rocks which form the sea-coast, the traveller cannot find sound footing anywhere (Byron, 1768, 112). Byron and his fellow survivors of the wreck kept life in their bodies mainly by capturing geese, especially the Canquén (Chlocphaga poliocephala), according to Crawshay (1907, 101). Relatively little is known about the pelagic bird life of this coast because, so far as I can learn, no one has undertaken collecting in the ocean just to westward, nor has any naturalist explored the difficult land areas in search of breeding grounds of petrels which are certain to be found here sooner or later. It is at about the latitude of the Guayaneco Islands that the character of the Magellanic coastal forest is transformed. Southward of this point the floor of the dense woods of evergreen beech is made up mostly of mosses, peat, etc. To northwards there is an impenetrable undergrowth of bamboos; the southern beeches are replaced by other species of the same genus, likewise evergreen; the diversity of trees increases, and the boles and limbs supply a hold for epi- phytes and lianas unknown nearer Cape Horn. In the words of Hellmayr (1932, 20), Skottsberg . . . divides the south-Chilean rain forest into two subsections: the ‘‘Valdivian,”’ richer in species and luxuriance, and the ‘‘Magellanic,’’ characterized by the predominance of Patagonian trees, notably Northofagus betulotdes. The dividing line is drawn along 48° S. lat., which also coincides with the southern limit of the range of certain species of birds. However, as the westerlies hinder the development of forest at sea level on the southwestern coasts of Tierra del Fuego, so do they above an altitude of 400 meters on the Peninsula of Taitao, and above 800 meters on Chiloé Island. Between the Guayaneco Islands and the nearly decapitated peninsula of Tres Montes is the Gulf of Pefias, the first great gap in the seaward chain of islands. TIERRA DEL FUEGO AND SOUTHERN CHILE 245 Bryce (1912, 289) thus records his impressions from a steamship along this stretch of Patagonia: Among the headlands which we saw along this stern and lofty coast, two were especially striking from their height and form. One is called Tres Montes. Heavy clouds hid its top, but two thousand feet were visible of the steep face that rose above the sea. Further south the huge tabular mass of Cape St. George, grand and grey in its drapery of mists, looked out over billows, the spray of whose crests as they broke upon the rocks could be seen fifteen miles away. There is not in the world a coast more terrible than this. No hope for a ship driven in against it by the strong currents and the resistless western swell. In a cove near Cape Tres Montes Darwin (1933, 262) noted vast numbers of seals and gulls on the rocks in January, 1835. He writes, too, of Turkey Vultures standing among the other creatures, as though patiently waiting for a seal to die. The vulture seems incongruous on this cool, soaked, and tumultuous coast, and yet it is a characteristic bird of the shore throughout the range of the southern sea-lion, upon the mortality of which it in part depends. While crossing the gulf from Cape Tres Montes, during a storm on July 4, 1914, Beck saw very large numbers of small petrels which he could not identify— an indication of the profitable work still to be accomplished through collecting in this region. In the northeastern extremity of the Gulf of Pefias, the little Rio San Tadeo reaches the sea, the southernmost stream on the west coast with its mouth close to the open ocean; all its neighbors enter deep, landlocked fiords and esteros. The San Tadeo is not navigable for seagoing vessels. It flows through lowlands, and its mouth is blocked by a great bar, with long sandy beaches on either side. Within, the river widens and deepens, and in freshet season it carries tree trunks down to the swamps near its mouth. Somewhere near its headwaters is the traditional portage across which the Indians formerly dragged their canoes from one shore to another of the Isthmus of Ofqui, and so made their way from the Gulf of Pefias to the Estero Elefantes and other waters to northward of the Peninsula of Taitao. To schooners and small trading steamers this peninsula is a barrier between waters; to southward is the domain of the port of Magallanes, to northward that of Chiloé. Here again a maze of numberless channels makes the ‘‘back country” acces- sible to vessels engaged in the lumber and fishing industries. The region has a romantic history, which is still largely to be written. With abundance of virgin territory among these wild islands, the number of settlements and their wide distribution are, nevertheless, surprising. Furthermore, the secluded harbors are so many and so secret, the turnings of the inlets so endless, that a pirate might lie at snug ease while all the navies afloat threaded the labyrinths in vain. I have no idea whether or not the Estero Elefantes is named for the long extinct sea-elephants that once ranged far northward along the west coast of South America, as the Patagonian sea-lion still does. Its smooth, beachless shores are steep on both sides, and above them forest-clad heights rise more than a thousand meters. Looking southward, the white ribbon of the San 246 OCEANIC BIRDS OF SOUTH AMERICA Rafael Glacier looms above the blind end of the gulch. To northward, at the opening of the estero, there are constant and dangerous tide-rips, and a heavy surf rolls across the mouth of the Rio Huemules despite the barrier of islands that lies between it and the open Pacific. Huge palisades of uprooted trees, thrown on the bar during floods, block the entrance, while inside the mouth the flat marshland is covered as far as the eye can reach with the dead wreckage of a forest. When the exits of smaller neighboring streams become choked by such submerged timber, their flood-waters burst the banks and carve new courses through the floor of coniferous woodlands (Steffen, 1929, 173). Only a little to northward, more or less navigable rivers begin, the first being the Aisen, which flows from one of the great watersheds of the Andes and has many settlements in its valley lands. Still more available as a trans- portation route for small craft is the Rio Palena, one of the most important rivers in the southern part of a country which, in a sense, has no rivers deserving that adjective. Within this statement, incidentally, is encompassed one of the most profound geographic differences between the eastern and western sea- boards of the continent. North of the almost insular block of Taitao, the island fringe continues as the jumble of the Chonos Archipelago and the Guaitecas, the latter abutting on the Gulf of Corcovado, the second wide gap in the chain. Throughout this multitude of islands, those nearest the Pacific, which receive the full force of the west wind, tend to be bare, while those toward the mainland are covered with luxuriant tree and thicket growth from the water's edge. However, the relatively isolated, offshore islands of Guamblin and Guafo, about which regrettably little is known, are said to be covered with dense vegetation, in- cluding excellent ‘“‘building timber.’’ The pilot books hint obscurely of wild dogs, and even of wild men, on Guafo! When Dr. Chapman approached the northern shores of the Guaitecas Islands, from Chiloé, on December 31, 1923, one member of his party bet another a Chilean dollar that there would be a Steamer Duck upon the first rock sighted. The bird was there as though by appointment. Entering among the islands, Chapman (1933, 360) writes, We were in a world of mountain tops flooded by the sea; there were little islands and large ones, low islands and high ones. Symmetrically rounded, in profile they were fluted, crenate, notched or dented but never peaked or pointed. Nearly every one, except the smallest rocky islet, was wooded from high-water mark to summit by a forest so thickly grown with trees and undergrowth, so clogged with fallen limbs and logs, that one could enter it only on hands and knees and then only for a yard or two. I was reminded of Darwin's story of the sailors from the Beagle (perhaps our Captain's father was among them) who, unable to penetrate the vegetation of a mountain side, clambered over it, stopping now and then to “‘take soundings.” After writing of the Steamer Ducks, Kelp Geese, and penguins of the Guaite- cas, the same authority comments upon the land birds strangely grouped with these sub-antarctic sea fowl. The parrots, for which the forest seemed so well adapted, were actually represented by a flock of Chilean Paroquets that flew screaming through the tree tops. There was a Magellan Wood- pecker, a superb black bird, fifteen inches long with a flaming scarlet crest, and many Ruby- TIERRA DEL FUEGO AND SOUTHERN CHILE 247 crowned Hummingbirds in the coihue blossoms. Penguins, Parrots, Woodpeckers, Humming- birds, there is an avian association rivaling the fauna of the Swiss Family Robinson! Nor can I do better than to continue to accompany the same master of South American ornithology and zodgeography from the Guaitecas to the opposite continental shore. January 6th, we steamed across the Gulf of Corcovado and, after a six-hour voyage, reached the mouth of the Rio Rodriguez. Here we replenished our supply of fresh water from a stream cascading down the mountain, picked strawberries on the shore, and entered the most beautiful forest growth we saw in southern Chile. Thence we continued up the eastern or mainland side of the gulf, exploring rivers, bays and fiords at will, always finding something new and interesting, always in the presence of supremely beautiful scenery. Corcovado was the dominant mountain, and a bay in which the German warship Dresden had sought seclusion had acquired historic distinction. . . . Assured by Captain Yates that we would find Black-necked Swans in the Bay of Pumalin [in the shelter of Llahuen Island], we entered this land-locked body of water and found not only Swans but a place so filled with birds of land and water that of all the spots I have visited in Chile this is the one to which I long most to return. There were pastures at the head of the bay, evidently on an alluvial fan through which a river wound its way across the mud flats at its mouth. Cattle had been raised here and a ranch house was occupied. . . . Hudsonian Curlews called through the night, and in the morning we found them with Oyster- catchers on the mud flats. A flock of twenty Ashy-headed Geese were asleep on a sandbar, and on an adjoining pasture seven Ibises and four or fivé hundred Dominican Gulls were feeding. On the river there were Teals (Nertion flavirostre) and broods of downy Widgeons attended by both parents. Penguins and Quettros rested on the pebbly bay shore, and on the bay itself there was a fleet of nearly one hundred and fifty of the Swans we had come to see. Singly or in pairs, they passed my hiding-place, their white bodies punctuated by their black necks and red bills. A superb picture they made breasting the crisp waves, with snowy Corcovado rising from the dark forest wall behind them... . Land, as well as water, birds were abundant at Pumalin. One of our two days there I followed the course of the little river through the pasture to the woods. Thrushes, House Wrens, Swallows, Goldfinches, Andean White-throated Sparrows and Kingfishers, all resembling our North American species in voice, appearance, or both, bespoke the relations of the South and North Temperate Zones. But Cinclodes (Cinclodes fuscus), the Spinetail (Aphrastura), the Babbler (Scytalopus), the Turco (Hylactes) and the Cheucau (Preroptochus) told with equal force of a distinct South Temperate fauna. If we knew of Chiloé only what Darwin wrote in his field journal a hundred years ago, we should still be relatively up to date so far as concerns information within the field of natural history. An island 180 kilometers in length, it is largely covered with impenetrable forest as a result of the heavy precipitation. The soil is so wet that roads avail little. In earlier times persons have lost their lives in the mud, and in 1856 a newspaper reported that cattle stealing was unknown inasmuch as no thief could escape with such booty through the damp woods. Most of the trees in this forest have little economic value. The coihue CNothofagus dombeyi) is too weak for lumber and too wet to burn. It simply keeps out the sun with its beautiful foliage and makes a quagmire of the ground. The alerce (Fitzroya patagonica) is, however, one of the useful giant conifers of America, sometimes 60 meters in height and nearly 5 in diameter. Trees of this species over 2000 years old are well authenticated. The wood is used in large quantities for shingles (Rudolph, 1929, 61). 248 OCEANIC BIRDS OF SOUTH AMERICA Chiloé is lower than most of the large islands to southward of it, although the barely separated Isla de San Pedro, at its southeastern corner, attains an altitude of 975 meters. Its channels are also much shallower, and the easy gradient of the land sloping under the sea produces extensive mud flats, exposed at low tide. Owing to the relations of shore and water, the tidal rise is greater in the gulfs of Ancid and Corcovado than elsewhere along the southern west coast. In places it amounts to 7 meters; tidal whirls, known as “‘rayas,”’ are notorious in Reloncavi Sound, at the head of the Gulf of Anc&d, and terrific rapids rush back and forth through the Chacao Narrows, which divide the northern end of Chiloé from terra firma. All of the characteristics cited are favorable to teeming littoral life; the shoals, the currents, the extensive fields of kelp, the profusion of fish and shellfish, the many streams and the wealth of interior lakes and morasses, combine to make the island a paradise of water- fowl. On the flats off Anctd, during early April, 1914, Beck saw enormous hordes of migrant shore birds from North America, including one flock of Sanderlings (Crocethia alba) numbering 5000, more or less. As the southern- hemisphere autumn advanced and the visiting species departed, their place was taken by flamingoes and other waders from breeding grounds close at hand. To Darwin (1933, 231) the Chiloé country seemed much like Tierra del Fuego, but with incomparably more beautiful woodland. Instead of dusky uniformity, he found almost the quality of a tropical environment, particularly in the equability of temperature and the presence of tall arborescent bamboos. A calm day, with the snowy cones of the cordillera showing over an inland sea of glassy smoothness, rippled only here and there by a porpoise or a Loggerhead Duck—such a scene framed by ferns and parasitic plants and the trunks of sweet- smelling evergreens, enhanced the tropical illusion. However, the terrible surf that pounded the seaward coast after westerly gales reminded him sharply again of the cruel reaches tosouthward. The roar of such waves, according to Darwin, can be heard at night in Castro, across 35 kilometers of hilly wilderness. Darwin also reports the attempt of a party from the ‘Beagle’ to reach the summit of San Pedro. The wood was so intricate, with its mass of dead and dying trunks, that for a quarter of an hour at a time the feet of the climbers never touched the ground, being generally two or three times a man’s height above it. In other places they had to crawl like foxes, one after the other, under the confusion of rotten logs. Darwin remarks that the members of the party were like fish struggling ina net. At an elevation of 300 meters, or thereabouts, they found stunted examples of the beeches which, nearer the Strait of Magellan, grow at sea level. Despite the picture thus far given, Chiloé is a seat of ancient human settle- ment. The eastern shore, along the gulf, is relatively populous, and long culti- vation has given parts of the region an aspect reminiscent of islands in the Baltic. Farm and pasture have been burnt out of the forest. There are numbers of villages, ‘each with its little white church. To look from . . . pastoral slopes across the Gulf of Corcovado to the snowy wall of the Andes is like seeing the Rockies from New England’’ (Chapman, 1933, 360). TIERRA DEL FUEGO AND SOUTHERN CHILE 249 The climate of Chiloé is altogether characteristic of the west coast of South America, where mean annual temperatures tend to lower only slowly toward the south. The southern border of the Pacific high-pressure area oscillates seasonally, being a little to southward of its average latitude in summer and a little to northward in winter. Consequently the system of northerly winds extends during the winter as far north as Coquimbo, while in summer the south- erly winds reach southward to Chiloé or even to the Gulf of Pefias. This means, however, that while winter temperatures remain approximately the same throughout ten degrees of latitude (as at Valparaiso and Ancid), the summer, on the contrary, is not only more humid toward the south, but also much colder. The average temperature for the warmest month of the year is 19.9° C. at Santiago, but only 14.5° C. at Puerto Montt. Furthermore, winter temperatures tend to be slightly higher than in corresponding latitudes on the Atlantic coast, while midsummer (January) temperatures are markedly lower. It is less the lowering of the mean temperature than the lowering of the summer temperature which controls the distribution of many forms of bird life and likewise deter- mines the southern limit of agriculture. In Chubat, Argentina, the wheat harvest is assured, while in Chiloé grain must be mowed before it is mature. The same meteorological factors in part control the asymmetrical distribution of coastal birds, such as geese, steamer ducks, shags, and oyster-catchers in southern South America. It is not surprising to find at Chiloé the approximate northern limit of a number of forms which, on our southward journey down the Atlantic coast we encountered for the first time in the higher latitude of the Falkland Islands and southern Patagonia. If one can credit the pilot books, Chiloé is an island of picturesque weather phenomena. Waterspouts are frequently formed along the gulf coast. Thunder is rarely heard at sea, but is common above the forest, preceded by spark-like flashes of lightning. After brief periods of clear skies and calm air, the atmos- phere toward the continent sometimes becomes peculiarly transparent, with a light green tint in the sky. Refractive stratification of the air causes promon- tories, rocks, islets, and houses to take on vibratory movements which distort or divide them, and to rise high or disappear in mirage. At such times visibility is often abnormally increased, so that details of trees and other features can be distinguished at a great distance. Then, as a rule, coffee-hued cumulus clouds begin to gather in the northerly sky, soon piling up and darkening into nimbus; the spreading curtain dims the scene to obscurity and the floodgates of heaven open with the tempest. c. The Gulf of Coronados to Valparaiso. The shores of the Gulf of Coronados, just north of Chiloé, are low where the Rio Maullin meanders to the Pacific from Lake Llanquihue, but they soon rise to a high and rugged coast with timbered seaward slopes. Short streams are numerous, as are tiny islets or farallones, but the coast is exposed, and pro- tected roadsteads are scarce. Even the mouth of a substantial river, such as the Bueno, is blocked off completely for navigational purposes from the ocean. 250 OCEANIC BIRDS OF SOUTH AMERICA Of the many brimming rivers that sweep down from the Andes across the Central Valley none is more beautiful in its lower course than is the Rio Bueno. It has in the course of ages cloven for itself through the hard rocks of the Coast Range a channel so deep that the tide comes up to the little town of Trumajo forty miles from the sea, and from that town small steamers can pass all the way to the bar at its mouth. In one of these little craft . . . wespenta long day in sailing down and back again. The hills on each side, sometimes hanging steeply over the stream, some- times receding where a narrow glen opened, were clothed with the richest wood. It was a bril- liant day in October, answering to our April, and the sun brought out an infinite variety of shades of green in the young foliage in these glens, the trees all new to us, and the spaces between them filled with climbing plants hanging in festoons from the boughs. Wild ducks and other water- birds fluttered over the water and rose in flocks as the little vessel moved onward, and green paro- quets called from the thickets. As it nears the sea, the river spreads into a wide deep pool under a crescent of bold cliffs, and at the end of this is seen the bar, a stretch of sand on which the huge rollers of the Pacific break in foam. There is a lighthouse and a few houses near a flat stretch of meadow by the banks, the grass as green and the flowers as abundant as in Ireland. Specially vivid were the yellow masses of gorse, apparently the same species as our own, and, if possible, even more profuse in its blossoms than on those Cornish shores of which it is the chief ornament. I have seen few bits of coast more picturesque than this meeting of the still, dark river and the flashing spray of ocean under rocks clothed with feathery woods (Bryce, 1912, 242). Galera Point, which is heavily wooded, marks the salient at which the coast turns northeastward toward Puerto de Corral, at the common mouth of several rivers. This is the seaport of Valdivia, which lies upstream hidden in a wood of apple trees. In this neighborhood the evergreens are considerably less numerous than at Chiloé, and the native forest is of a brighter, more lively green. From Corral, Beck undertook important collecting during the Brewster- Sanford Expedition, making the city his last continental base until he reached Magallanes in the Strait of Magellan. He found the neighborhood of the whal- ing station at San Carlos to be particularly good hunting ground. Here, during September and October, 1913, he took many migrants from the south, such as Royal, Wandering, and Black-browed Albatrosses, Giant and Silver-gray Ful- mars, the Cape Pigeon, Shoemaker (Procellaria), the Fuegian race of Wilson's Petrel, and Sooty Shearwater. Among visitors from the north, taken in the same waters and season, were the Peruvian Pelican, Cormorant, and Booby, the Gray Gull (Larus modestus), as well as the Red Phalarope and two species of jaegers from the northern hemisphere. More or less local species, mingling among those of the other two groups, included the Magellanic Penguin, the Bigiia, Red-footed, and Magellanic Shags, the Pink-footed Shearwater, Peruvian Diving Petrel, Cinnamon Skua, Kelp Gull, and South American Tern. To northward of Corral the range of hills recedes from the ocean, and shoal and sandy shores, with occasional cliffs, lie in front of a broad fertile plain. Between Punta de Nihue and the latitude of Mocha Island a heavy surf breaks even in fine weather. Many areas of fresh water famous for their wild fowl, such as the extensive Laguna del Budi, nestle between the sea beach and the foothills, and some of the streams are broader and more impressive along their middle reaches than near their mouths. Isla Mocha, which lies in latitude 38° 20’ S., 33 kilometers off Punta Tirta, is one of the latest of the Chilean islands to be visited by an ornithologist. In TIERRA DEL FUEGO AND SOUTHERN CHILE 251 1932, Mr. D. S. Bullock, who has had Jong associations with the American Museum of Natural History, made a reconnaissance at Mocha, and has sent us a description of the island from which most of the following data are derived, as well as an excellent series of the resident birds. Mocha is about 11 kilometers in length, with an area of 50 square kilometers, of which about half is mountainous and covered with virgin for- est. The slopes are in many places so steep that it is impossible to climb them; the summit, Cerro Colmenares, has an altitude of 539 meters. Encircling the greater part of the high central mass is a more or less level plain devoted to graz- ing and agriculture, the best terrain for the latter purpose being close under the foot of the hills on the northerly side. Farm and pasture land, and tracts of marsh, are crossed by numerous brooks, none of which can be regarded as perennial. Nevertheless, their lower courses remain damp, and support breeding sites for such birds as fre- quent the lowlands. The hills, with their woods and thickets, supply a totally different habitat, occupied mainly by a restricted avifauna. The Mocha forest is made up largely of tique (Aextoxicum punctatum), intermingled in places with laurel (Laurelia aromatica), lingue (Persea lingue) and boldo (Boldoa fragrans), of which the last is most widely distributed. On the lower slopes there is much underbrush, in which the chilco (Fuchsia coccinea), the common wild fuchsia of southern Chile, is conspicuous. Whitethorn (Rhaphithalmus cyanocarpus) and maqui (Aristotelia maqui) also enter into the formation of thickets. The latter of these is able to germinate and thrive in the deepest shadow. It sprouts from the base of its stem with amazing speed. Its berries are the staple food of the native thrush of Mocha, which sows them far and wide. Several species of myrtles also flourish in the swamps and even among the hills. The coast presents an environment suited to Beck's route -—-— Chapin’s route I.Mas Afuera| G. de Arauco. 1. St? Maria DER Bio-Bio Valdivia 40 6 del Corcovado 1.$.Pedro R Palena ho G. de Perias Arc Guayanecost inal Macias \.Campana/)3 |.Wellington @ Pte Julian, |. Mornington. Fic. 38. Southern Pacific coast of South America and the asso- ciated islands. For detailed maps of the latter see figures 39 (Juan Fernandez), 40 (San Felix and San Ambrosio). Beck's course and local collecting stations, and Chapin’s voyage with Crocker in the “Zaca’ are indicated. Local field work of the American Museum is shown by figures in circles: 6, Chapman, Walcott, and Percy, at the Guaitecas Islands and in the Gulf of Corcovado; 7, Bullock, at Toltén, Mocha Is- land, and the Gulf of Arauco; 8, Hallinan, north of Coquimbo. many oceanic and beach birds. In certain places sandy areas with sparse vege- tation afford nesting grounds to the Snowy and Falkland Plovers and to oyster-catchers. Rocky promontories are the resort of gulls and Red-footed Cormorants, while Blue-eyed Shags (Phalacrocorax atriceps) nest on small flat- 252 OCEANIC BIRDS OF SOUTH AMERICA topped islets just offshore. Isla Muerta is occupied by a dense colony of Cholos or Peruvian Cormorants (Phalacrocorax bougainvilliz), which are here at their southernmost breeding station. The Isla de las Docas, being overgrown with a tangle of wild mustard, offers ideal nesting sites for the Magellanic Penguins. Shearwaters burrow into the forest floor on the slopes and heights of the main island. The habitats, as will be seen, are highly diversified, and Mr. Bullock's list of birds comprises about 85 species, some of which are only transients (Chapman, 1934, 1). A few of the native passerine forms are endemic races. Certain birds very common on the opposite mainland are, however, wanting at Mocha, and for no understandable reason. Among these are three species of woodpeckers, a mockingbird, and the Loica (Pegites). Visiting sea birds to the waters about Mocha include the Black-browed Albatross, Giant Fulmar, Cape Pigeon, Peruvian Booby, Cinnamon Skua, and Magellanic Gull (Leucophaeus scoresbit). Among waterfowl and waders of the pools, streams, and marshes, Mr. Bullock lists grebes, rails, gallinules, the Pintail Duck, flamingoes, ibises, and night herons, as well as many limicoline species, both resident and migrant. The breeding sea birds and littoral birds, of the groups to which this book gives particular recognition, are the following. PENGUINS CoRrMORANTS Spheniscus magellanicus Phalacrocorax olivaceus Phalacrocorax gaimardi Phalacrocorax bougainvillit Phalacrocorax atriceps SHEARWATERS Puffinus creatopus Puffinus griseus (probably a breeder) OysTER-CATCHERS Haematopus ostralegus pitanay PELICANS Haematopus ater Pelecanus occidentalis thagus Gir Larus dominicanus There are few other parts of South America in which four species of cormo- rants can be found nesting at a single station. In the latitude of Mocha, blue skies may be said to begin. Mere fogs and mists here tend to supplant in part the pounding rains of Chiloé, and the con- tinental mainland is less thickly concealed under trees. We have at length a suggestion of the pronounced stages of climatic space-succession which are further expressed at the Rio Bio-Bio (37° S.), near which the distinctive char- acter of the sub-antarctic forest ends, and at the Rio Limari (30° 45’ S.), where the woods are reduced to patches and the spell of the desert is foretold. Northward from Mocha a more rugged coast, with many submerged rocks and a series of headlands, leads to Punta Lavarié, the tip of a great triangular peninsula which, with Santa Maria Island, shelters the Gulf of Arauco and its ports of Coronel and Lota. Santa Maria, although inhabited and well cultivated, supports many sea birds on its promontories and farallones. Here Mr. Bullock TIERRA DEL FUEGO AND SOUTHERN CHILE 253 saw flocks of the sub-antarctic Blue-eyed Shag (Phalacrocorax atriceps) during the height of its breeding season, a northern record for this bird. Bryce (1912, 226) thus describes the eastern side of the gulf: The shore, sometimes rocky, sometimes bordered by thickets or grassy flats behind the beaches, is extremely picturesque; and were it in the populous parts of Europe or North America, it would be lined by summer cottages and alive with children. But its vegetation and general aspect are curiously unlike those of the Atlantic coasts of either of those two continents, and remind one rather of California. At Lota, the hills rise boldly from the sea and a large island [Santa Maria] lying some way out gives variety to the ocean view. The Bio-Bio, which flows into the northern end of the gulf, is not accessible from the ocean, and Concepcién, a short distance upstream, finds its outlet through Talcaguano on the bay to northward. Here the Peninsula de Tumbes on the west, and the island of Quiriquina in the mouth, more or less repeat the form of the larger Gulf of Arauco. Coppinger (1884, 96) gives the following picture of the natural history of the bay as he found it in the year 1879. There was a long, low sandy island (Isla de los Reyes) lying across the head of Talcahuano Bay, and inhabited only by a couple of shepherds who were looking after a herd of cattle and horses. . . . On the mainland immediately adjoining the island, I found a great marshy plain of many miles in extent, and intersected in various directions by deep muddy ditches which com- municated with the sea, and at high tide brought supplies of sea water to a chain of broad, shallow lagoons, the home of multitudes of waterfowl. Pin-tail ducks, widgeon, herons, curlew, flamin- goes, turkey-buzzards, gulls, lapwings, and sandpipers found here a congenial home. The shrill, harsh cry of the spur-winged lapwing (the ‘“‘terotero’’ of the Pampas so graphically described by Darwin) was forever scaring the other peacefully-disposed birds, and at the same time invoking maledictions from the sportsman. When the first ebb of the tide left bare the mudbanks in the lagoons, the gulls and curlews collected in vast numbers for their diurnal meal. Of the gulls only three kinds were seen, viz., L. Dominicanus, L. Glaucodes, and L. Maculipennis. The latter were in various conditions of plumage; some birds haying a deep black hood, and others with a head almost entirely white, while between these two extremes, there was every gradation. The turkey-buzzards derived a plentiful supply of food from the bodies of fish stranded on the beach. For some reason or other dog-fish were constantly coming to grief in this way, bodies of fish, two and three feet long, being met with sometimes, all along the beach, at average distances of about one hundred yards apart. From this point a long stretch of coast without marked indentations, and with no islands other than small outlying rocks, continues its slightly easterly trend toward Valparaiso. Many rivers, of progressively reduced volume north- ward, enter the ocean, and here and there are standing waters behind the shore, such as the great lagoon of Vichuquen in latitude 34° 50’S. Eigenmann (1927, pl. 1) has published graphic charts that show the lake and river systems of central Chile. South of about latitude 28° S., the coast is seen to be well covered with the dendritic designs of streams that flow for at least a part of each year. North of the rivers Guasco and Copiapé, however, an enormously long desert coast is broken only by the Rio Loa. Approaching Valparaiso from the south, one passes a rather desolate shore lined with flattish hills that are white with fossil shell-lime. Vegetation is scarce at most seasons, and algarrobo trees bespeak a climate far removed from that of Valdivia, although throughout the winter (April to October) there is 254 OCEANIC BIRDS OF SOUTH AMERICA more or less transitory verdure near the shore, supported by the rains brought by northerly winds. We are here north of the zone of permanent westerlies. The summer is a time of southerly and southwesterly winds, while the winter has variable winds and tempests from the north. There is also a daily alternating system in summer, the “‘travesia’’ or southerly wind beginning in the forenoon and blowing until evening, to be replaced then by the ‘‘terral’’ or wind from the heights. The latter is the cold breeze that produces the morning mists. Temperatures at the coast continue equable, as farther southward; the difference in the means of the warmest and coldest months at Valparaiso is only 6°C. The mean annual temperature is 2.1° lower than that of Montevideo, on the opposite coast of the continent. Like most of the middle Chilean ports, Valparaiso is situated to northward of projecting points which furnish full shelter against all but northerly winds. The latter sometimes wreak much damage. It is usually calm in the bay during the morning, even when a strong breeze prevails out at sea but, on the other hand, the harbor is subject to violent southeast squalls from the hills when the southerly winds offshore are only moderate. Both of these phenomena seem to be characteristic of bays or coves lying below high hills from this point north- ward through Peru. The harbor of Valparaiso is at times a marvelous resort of waterfowl, being often filled with combined companies of guano birds from the north and petrels or other wanderers from the south. Certain species, such as Skimmers (Rynchops) and Black-necked Swans, reported there in great flights a century ago (von Tschudi, 1846, 1, 35) might nowadays, however, be sought in vain. To northward of the bay are picturesque rock formations that have evidently been elevated only recently from lower levels by the frequent seismic readjust- ments of this coast. Headlands are numerous, but there is none that shelters another great niche in the coast, like the bays of Arauco, Talcaguano, and Valparaiso, until we pass the mouth of the Rio Limari and reach the peninsula called La Lengua de Vaca, which stands to windward of the bays of Tongoi and Coquimbo. ? 10. Tue Paciric Sus-Tropicat IsLaANDs a. Juan Fernandez. The Juan Fernandez Islands are in a latitude slightly south of that of Val- paraiso, the inner island, Mas Atierra, being 667 kilometers from the continent. Mas Afuera is 167 kilometers farther west and a very little to southward. Mas Atierra is 22 kilometers in length, narrow, bent like a boomerang, and reaches an altitude of 916 meters. Off its southwestern tip lies the small, waterless island of Santa Clara. Mas Afuera is smaller than the main island, but higher (1650 meters), of roughly rectangular outline, and cut by very deep, east-west, gorge-like valleys. The latter island has been during recent years uninhabited. Mas Atierra is the seat of a fishery for clawless lobsters (Palinurus), which is responsible for regular communication with the mainland. Many voyagers, such as Scouler (1826, 205) and von Tschudi (1846, 1, 38), have likened Juan Fernandez, in form, climate, and the superficial appearance PACIFIC SUB-TROPICAL ISLANDS 255 JUAN FERNANDEZ ISLANDS Miles 30 Kilometers 0510 20 % 40 — oo El : ey one t 33°30 Heights in meters tu See SOT G00 ind Bay} North Pt Pao npe! 10" £ . Island Pt, te, Mas Atierra I. ; i ballena P Mas Afuera | is 5t8 Clara iP aS) a, feiss allena } A Colonia Roodstead ral See ian Bae Los Inocentes Lg aa i] 0" Negra pt Fic. 39., of the original vegetation, to Madeira, or to Flores of the Azores. The islands are built of greatly eroded, Tertiary, volcanic rocks, and from the wild peaks and crests forest-clad valleys descend toward the coast on one or more sides. Many of the seaward faces are, however, extremely precipitous; on the west, Mas Afuera, for example, drops away in a sheer steep of 1500 meters, more or less. The climate of Juan Fernandez is moister and milder than that of the oppo- site coast of Chile. The islands feel the influence of the warm circulation of the Mentor Current rather more than that of the cooler coastal water. The mean annual temperature at Mas Atierra is 15.5° C., the August and February averages being, respectively, 12.4° and 19°. Rainfall amounts to 1008.9 millimeters per annum. It is least in January and heaviest in June, and at all seasons is concen- trated particularly on the higher eastern and central portions of Mas Atterra, where the air-currents condense their moisture after rising suddenly from the southerly shore. There are no adequate meteorological data from Mas Afuera. The climate, the nature of the vegetation, and doubtless that of the animal life, cannot be discussed entirely in terms of conditions at sea level; the situation is complicated by altitude. The lofty upland of Mas Afuera, for instance, is an alpine tundra where, in 1908, Skottsberg (1918, 362) discovered a totally unsuspected Magellanic flora comprising such plants as a sub-antarctic bramble (Rubus geoides), a club moss (Lycopodium magellanicum), and a mountain aster (Lagenophora). Yet, as regards sea birds, it is apparently the temperature of the ocean surface, rather than of the terrestrial nesting site, which exercises full control. At Mas Afuera, which is the closer of the islands to warm oceanic regions, the petrel (Prerodroma leucoptera masafuerae) that breeds only on the coldest, sometimes frost-covered heights, is a bird of definitely pan-tropical affinities. At Mas Atierra and Santa Clara, which are nearer to cool Humboldt Current influences, this petrel is unknown, and is replaced by Prerodroma cookii defilippiana, a bird of sub-antarctic affinities which, notwithstanding, nests at much lower and warmer altitudes than the petrel of Mas Afuera. The flora of Juan Fernandez has been studied by Skottsberg, who finds on one or another of the islands both a sub-antarctic and a tropical American 256 OCEANIC BIRDS OF SOUTH AMERICA element, as well as a strange endemic moiety which offers only doubtful hints of relationship with plants in distant parts of the Pacific. The vertical zones comprise forest, grassy meadows and fern-beds, and mountain heath, with large sterile areas on the precipices and unfavorable exposures. The forests, which ascend to 700 meters or thereabouts, are of sub-tropical type, with evergreen forms predominating; lianas are Jacking, their place being taken by ferns with winding rhizomes reaching high up the trunks of the trees. Myrtles, tree ferns (Dicksonia), and a large endemic chonta palm (Juania australis) are characteris- tic. The native sandalwood, known nowhere else in the world, has been extir- pated, and weeds and continental thicket-plants are winning a slow victory over the aboriginal flora. Toward the upper limit of the dense forest are a number of peculiar miniature trees of endemic genera, belonging to ordinarily non-arborescent families. Some of these also flourish on the grass-grown steppe of Santa Clara. Goetsch (1933, 26) lists Collembola, ichneumons, termites, Neuroptera, Hemiptera, grasshoppers, beetles, and Lepidoptera among the insects of Juan Fernandez. Many species, he states, are endemic but, except for cosmopolitan types, they are all related to South American forms. Fifteen of 31 species of land mollusks are endemic, most of these showing Polynesian relationships. There are neither reptiles nor amphibians, and no mammals other than the practically exterminated seals. Goetsch computes the percentage of endemism in the total fauna of the island, from lower invertebrates to birds, to be 44.5 per cent, although the figure can be taken only as a rough approximation. Both islands of Juan Fernandez were recently visited during the cruise of Mr. Templeton Crocker’s yacht “Zaca.’ The following notes are freely tran- scribed from the journal of my colleague, Dr. James P. Chapin, who was a member of the party. January 30, 1935. The southern side of Mas Afuera is very lofty and steep, with levels above a thousand meters hidden this morning in cloud. The cliffs do not drop so directly into the sea as at many of the Marquesas Islands, and there are a few grass-covered taluses. Above an altitude of about 600 meters the slopes at this end, even where nearly vertical, are clothed with green shrubbery and a few trees. The eastern side is less precipitous and more grassy. Goats are grazing on the lower hillsides. The recently abandoned penal settlement forms a group of about ten buildings at the mouth of a narrow gorge. One pine of some European species and a few large eucalyptus trees stand near the structures. Elsewhere the scanty vegetation gives clear evidence of relative drought at this season, which is the driest part of the whole year. On either side of the empty penal buildings—probably the only human colony ever established at Mas Afuera—trails zigzag up the steep hills. Behind, a road runs into the gorge, where the brook is dry save for a few stagnant pools. There are no woods in the floor of this valley, but small trees and large ferns grow rather thickly on the steep hillsides toward its head. Along the trail to northward of the buildings, the slope is grassy except where the turf and soil have been eroded down to the underlying lava, probably as a result of grass fires. At an alti- tude of about 400 meters are the first clumps of large ferns, resembling tree ferns without trunks: Above this level, trees with very small leaves also become numerous, and just below 500 meters is a gully filled with a forest of such trees, among which are mingled tree ferns standing five meters tall and with fronds measuring up to nearly three meters in length. There are also a few trees with acacia-like foliage, and here and there a maqui to which the native thrushes come for berries. PACIFIC SUB-TROPICAL ISLANDS 257 Still farther on, the road winds through woodland and then through more open grassy stretches, with scattered trees. In this belt, close to an altitude of 550 meters, are many burrows of Ptero- droma externa. The trail ends in a grove on the gradual slope, away from any ravine, and here is a rough shack presumably occupied at times by men from Mas Atierra who come to hunt goats. Above this altitude the slopes are more open and grassy, with many clumps of ferns, and with small patches of forest only in the ravines. In a few of the dampest spots at the bottoms of these, one sees the enormous leaves of Gunnera peltata which may have a length of three meters and a breadth of more than one. Above 700 meters and below 400, there are evidently no real woods on Mas Afuera, the forest zone therefore forming a fringe like the hair on a tonsured head. De- spite the aridity of the season, as shown by the parched grass and the state of the streams, none of the trees has shed its leaves. In the wooded belt, the thrush (Turdus falcklandii magellanicus) is the most common land bird. At lower and predominantly grassy parts of the island there are almost no birds except the hawk and, in the stream beds, the Cinclodes. January 31. As we approach Mas Atierra in early morning from its western end, it looks barren, as does also the same aspect of Santa Clara. There are patches of bare earth, much dry grass, a little shrubbery, but very few trees. Steep cliffs rise from the ocean. The top of Sugarloaf is capped with woods, in which five or six palms (Juania) are conspicuous. The lower parts of the valleys opening on the northerly shore are also barren but, after passing Selkirk's Cave, it becomes apparent that the bases of all the high central mountains along this side of the island are far more extensively forested than any corresponding part of Mas Afuera. The low ground near the village of San Juan Bautista seems parched, though many shade trees have been planted there. Large patches show where soil has been blown away by the wind. The road leading inland toward the base of the high peak called El Yunque is, however, well shaded by a low wood composed mainly of the maqui, introduced from the continent, which is steadily crowding back the native arborescent vegetation. At a distance of two and half kilometers from the shore, the present limit of the maqui is reached, and one enters a large level area with an altitude of about 300 meters which stretches toward the base of the mountains. Here is a small farm in the heart of a beautiful forest, with a goodly variety of indigenous trees, including a large Zanthoxylum reaching a height of 20 meters. This sylvan cover of this interior depression of Mas Atierra is not only far richer than that of Mas Afuera but it also extends very much nearer sea level. A richer land bird fauna reflects the condi- tion, for hummingbirds of two species are abundant, especially about a certain blue-flowered tree (Raphithamnus longiflorus). 1 am informed, too, that a barn owl, which has not been recorded in the literature, inhabits these forests. The large native Pterodroma, which is a different species from that of Mas Afuera, is said to nest only at altitudes above 500 meters. In swampy portions of the woodland are large patches of the remarkable Gunnera and many kinds of ferns. One of the latter climbs the tree trunks to a height of six or seven meters, and others are epiphytes among the branches above. A native cycad grows throughout the forest belt, and in the upper parts there are two or more species of tree ferns. According to Chapman (1934, 5) the indigenous land birds of Juan Fernandez number eight species, of which three are specifically and three others sub- specifically endemic. One hummingbird is represented by a distinct subspecies at each of the two main islands. The native buzzard (Buteo polyosoma exsul) is reported to fly regularly from its principal headquarters on Mas Afuera across the long stretch of ocean and back again (Lénnberg, 1921, 10), rarely remain- ing at the larger island for more than a day or two ata time. This hawk shares more or less of the tameness and lack of caution of its relatives at the Galapagos Islands, suggesting long isolation of its stock. Two introduced species flourish in a feral state, namely, the Rock Pigeon (Columba livia) and the California Quail (Lophortyx californica). The former nests on the cliffs, close to the native petrels. Casual visitors to the islands include both land and water birds from 258 OCEANIC BIRDS OF SOUTH AMERICA the continent, while the surrounding ocean is within the regular migratory range of many pelagic birds from the south, the islands of the Pacific to west- ward, and the northern hemisphere. The highly characteristic gulls, pelicans, cormorants, boobies, and diving petrels (Pelecanoides) of the Chilean cool- water littoral are, however, unknown at Juan Fernandez. The resident and breeding sea fowl are as follows: PENGUINS Spheniscus magellanicus PETRELS Puffinus creatopus Pterodroma neglecta Pterodroma cookti defilippiana Pterodroma externa Pterodroma leucoptera masafuerae Fregetta grallaria grallaria Of the above, only Pterodroma externa and P. 1. masafuerae are known certainly to breed at Mas Afuera. It is highly interesting that these two Procellariiformes are definitely sub-tropical members of the assemblage, while all of the other four species except P. neglecta have a certain sub-antarctic cast. During Mr. Beck’s field work in the neighborhood of Juan Fernandez, in December, 1913, he began to encounter examples of most of the resident sea birds when at a distance of 150 kilometers or thereabouts from the islands, although De Filippi’s Petrel was met with much nearer the continent. On the inter-island voyage, the most abundant species, by far, was the Pink-footed Shearwater. On his small boat trips offshore from Mas Atierra, he rarely found petrels of any sort within 5 kilometers of the shores. Among the transient sea birds known to have been taken on the Pacific within sight of Juan Fernandez are the following, species captured by Beck being marked with an asterisk: ‘‘Albatross,’’ ‘‘Gray-headed Albatross,’ *Diomedea exulans, Diomedea melan- ophris, *Macronectes giganteus, Daption capensis, Priocella antarctica, *Procellaria aequinoctialis, *Puffinus carneipes, Puffinus griseus, *Pterodroma cookii orientalis, Oceanites oceanicus, *Phalaropus fulicarius, Stercorarius sp.?, Sterna paradisaea, *Sterna fuscata. b. San Ambrosio and San Felix. These small islands lie on the same submarine ridge as Juan Fernandez, but 780 kilometers farther northward and about opposite Chafiaral on the Chilean coast. The ridge is defined by the 2000-meter line and to eastward is the Richards Deep, descending to 5000 meters. The islands therefore represent the summits of volcanoes comparable in height with the Andes. They are at a slightly greater distance from the nearest point of the continental shore than Mas Afuera, and are about 18 kilometers apart. San Ambrosio is an imposing rock of black basalt, rising 254 meters in a sheer precipice on its southern side and sloping to cliffs 100 meters high on the north. Landing is possible at one place, Covadonga Cove. The altitude of the PACIFIC SUB-TROPICAL ISLANDS 259 80°00" ss’ 50° SAN AMBROSIO AND SAN FELIX ISLES Miles 53 0 ' 4 é SS _oppr” Peterborough Cathedral Kilometers Felix Road 1) ai ea Ae dara eo \san Fotix ‘ “ Morro Amarillo Heights in meters () Gonzales I. 73 80°oo' Fic. 40. summit is 479 meters, and there the mists of adiabatic condensation support spreading shrubs and a few other plants. Dr. Ivan M. Johnston, of the Arnold Arboretum, writes me not only that the flora of San Ambrosio is considerably richer and more luxuriant than that of the lower island of San Felix, but that its green top is also the haunt of a number of insects and other native inverte- brates. San Felix, to northwestward, is a low crescentic platform. A rounded hill, the Cerro Amarillo, is 193 meters in altitude and of a cinnamon color. At the southeastern extremity is a separated hill of the same decomposed tuff, known as the Islota Gonzalez, and to northward is a single stack, the Roca Catedral or Peterborough Cathedral. The length of San Felix is about 3 kilometers, its width nowhere more than 1000 meters. The cliffs along the southern coast are of jet-black lava (Willis and Washington, 1924, 365). Little is known of the characteristics of the ocean about these islands, but their desolate aspect would indicate abnormally low temperatures for the lati- tude even at this considerable distance from the zone of upwelling along the South American coast. Precipitation must be scanty, judging not only by the relatively slight plant-growth but also by the fact that San Felix formerly yielded guano rich enough in nitrogen to warrant its transportation to Chile. The coasts of both islands abound with the same edible lobsters which support an industry at Juan Fernandez, and which are absent or scarce in corresponding latitudes of the mainland. After the Chilean earthquake of 1922, San Felix was visited by Professor Bailey Willis, whose report has been drawn upon above and who has courteously supplied me with some interesting notes on the natural history. He states that when Captain Stuart Campbell put in at the island for lobsters in October, 1922, its surface was covered with guano and the boobies were so numerous that the seamen ‘‘had to kick them out of the way."’ During the seismic dis- turbances of the following month, gases of volcanic origin issued from the crevices of San Felix and destroyed or drove away the greater part of the bird life. When Captain Campbell returned in February, 1923, there were only dead 260 OCEANIC BIRDS OF SOUTH AMERICA birds at the island, and he himself was overcome by gases upon his ascent to the basalt plateau. Likewise the spiny lobsters on the surrounding banks had been killed in large numbers, as reported by a diver who found the bottom strewn with their carapaces. Professor Willis arrived on a Chilean corvette in May, 1923. At that time a sulphurous odor and puffs of bluish vapor were still noticeable, but he satis- fied himself that there had been no active flow of lava from the crater of San Felix since the date of the original discovery of the island by Juan Fernandez in 1574. Small bushes were still growing on the old lava at the summit. Willis counted ‘‘twenty-five live birds and fifty dead ones on the entire island, where formerly there must have been many thousands.”’ My colleague, Dr. Chapin, who landed only at San Felix, collected there four species of plants, namely a sea-blite (Swaeda microphylla), a mallow (Cristaria insularis), a salt bush (Atriplex) of a new species, and a love-grass (Eragrostis) of a new species or new variety. The last is the first grass ever found at either of these islands; it seems to be a close relative of Eragrostis peruviana which thrives along the maritime ranges of the mainland from Lima southward to Taltal. Dr. Johnston, who has identified these specimens, informs me by letter that the plants of the two islands comprise about 14 species, of which all but three or four are apparently endemic, although the relationships of several are still dubious. With the exception of an endemic genus of composite (Tham- noseris), the affinities of the plants are clearly with elements in the flora of the coastal hills of Peru and Chile. Thamnoseris may perhaps have closer connec- tions with endemic genera of Juan Fernandez and Polynesia, and hence may be a representative of an ‘’Old Pacific’’ flora.* Until the recent cruise of the ‘Zaca,’ our information about the birds of San Felix and San Ambrosio had been woefully scant. The islands are an ornitho- logical station of particular interest because of their somewhat anomalous zonal position. They belong neither to the Humboldt Current nor yet to the fully tropical Pacific, and, while they share two or three of the native sea birds of Juan Fernandez, the avifauna has been alleged to have other components that would make an odd mixture. References to birds in the literature are disappointingly casual, and specimens have been extremely scarce and scattered. Morrell (1832, 117) states that ship’s provender in the form of eggs of sea birds may be gathered in any quantity at San Ambrosio during the months of Decem- ber and January. Those of the “‘shag and gannet,’’ he adds, are nearly equal in taste to hens’ eggs. It is highly improbable that any kind of shag or cormorant ever inhabited these islands. The presence of a gannet or booby has been known from photo- graphs of downy nestlings, and that of a Gray Ternlet (Procelsterna) and a petrel (Pterodroma cookii defilippiana) from one or two specimens of each. The native *Dr. Johnston has since written me that the Eragrostis proves to be a continental form, that Thamnoseris is represented by distinct species at San Felix and San Ambrosio, respectively, and that the affinities of the latter genus are American. Its relationship with the Polynesian Fétchza is illusory, and its only known close kin is Dendroseris, of Juan Fernandez. The important point is that the San Felix and San Ambrosio flora is definitely and entirely American, though with a remarkably high degree of endemism as to species and genera. PACIFIC SUB-TROPICAL ISLANDS 261 birds, which are all sea fowl, are now determinable from comprehensive col- lections made by Dr. Chapin and Mr. Jaques of the “Zaca’ party. They are as follows: PETRELS TERNS Pterodroma neglecta Sterna fuscata Pterodroma cookii defilippiana Procelsterna albivitta Fregetta grallaria grallaria Anotis stolidus Boostgs Sula dactylatra The assemblage is, in the main, sub-tropical but, in view of the presence of the Masked Booby, Sooty Tern, and Brown Noddy, it is somewhat surprising not to find other boobies, man-o’-war birds, and fairy terns (Gygis). The islands are doubtless to be regarded as a sub-tropical outpost at the verge of the cool littoral water. There is not a single resident species suggestive of the Humboldt Current avifauna. There remains for disposition the apparently well-substantiated record for San Ambrosio of a nesting colony of an American ‘‘marsh tern,”’ Sterna trudeaui (Bent, 1921, 228). The account is fully specific as to the discovery of these birds, and the collection of numerous eggs, at this island on December 17, 1907. Mr. Crocker’s party saw no trace of the species in February, 1935. It is hardly necessary to cite such negative testimony as evidence, however, for the zodgeographic probabilities are overwhelmingly against the presence of breed- ing Trudeau's Tern at these islands at any time of year. The record is doubtless based upon either an error or a hoax; if the former, it is possible that “San Am- brosio"’ refers to another locality in Chile, or that the collector's eggs of Sooty Terns or Noddies had become transposed with Trudeau's Tern eggs from the mainland. Both San Felix and San Ambrosio, of course, need thorough inspection in search of the still undiscovered breeding grounds of a number of petrels known only from waters off the Chilean and Peruvian coasts. The following account is freely transcribed from the notebook of Dr. James P. Chapin, who visited San Felix and San Ambrosio in the ‘Zaca’ as the guest of Mr. Templeton Crocker. February 18, 1935. As we approach Gonzales Island from the southeast, both it and San Felix seem wholly bare, but a nearer view shows small clumps of pale green plants on some of the slopes. There is no visible patch of grass, and nothing approaching the dignity of a tree or even a tall bush. The terrain appears more barren than any at the Galapagos except that of young lava‘fields. The western end of San Felix is a hill of soft yellow-brown volcanic rock and the lower part is made up of brownish-black lava cut into a sea cliff. An old cable hanging from the rock above the land- ing place enables one to reach the summit of the island, where there is a little soil resulting from the disintegration of the volcanic rock and the droppings of sea birds. Nearly the only vegetation consists of widely scattered plants with small succulent leaves, which turn from green to dark reddish as they age, so that the whole shrub often has a purplish appearance (Suaeda). Radiating from a central root, some of these plants are more than a meter in diameter, but they average about half that, with a height of 25 centimeters. Less common is a small, spiky, green herb, most examples of which prove to be completely dried out (Cristaria), and 262 OCEANIC BIRDS OF SOUTH AMERICA in one spot are numerous little grass plants pressed against the ground and long since desiccated . (Eragrostis). A fourth species grows on and about the Cerro Amarillo, at the western end of the island. This is the one first observed from sea. It has small gray-green leaves and tiny green- ish or yellowish flowers (Atriplex). It, too, radiates from a central root, reaches a diameter of upwards of a meter, and its woody stems are very brittle. Many feathers are entangled in all the plant clumps. All down the gentle slopes on the northern side of San Felix, flocks of sooty terns rise and fill the air, leaving on the ground, however, a few fledglings not developed quite enough to fly. The only. traces of eggs are broken shells and a few ancient, infertile examples. Old booby nests are recognizable by the radiating streaks of white, but only one occupied site has been seen. On this two birds are squatting, one of them attempting to incubate a thoroughly addled egg. The only present bird of the island that could have laid down such guano deposits as have been described would appear to be this species. Yet Sula dactylatra is at best never a truly colonial booby, such as Sula variegata of the Peruvian coastal islands. Even under conditions of maximum population, its nests tend to be somewhat widely scattered, like those of Sula nebouxii, rather than compactly placed like those of typical guano-producing sea fowl. On small piles of rock, especially in slight hollows, the gray ternlets are alighting freely, though no eggs or young are to be found. Pterodroma cookii defilippiana evidently nests in small cavities of the rock, particularly on the slopes of the Cerro Amarillo, where the soft stone weathers in such a manner as to provide innumerable little niches. In some such cavities there is gray down from nestlings that have departed, and in others the skeletons of birds that have died. No chicks are in the holes, but one has been shot on the water near shore and a second captured ona flat ledge above our landing place. A few adults are flying about over the ocean, and others along the high cliff that forms the southern edge of San Felix. One has even flown into a large cave close to the landing place. Approaching San Ambrosio from the westward, in afternoon, we can see that its lofty but rather level summit has considerably more vegetation than any part of San Felix. Certain plants appear to resemble the gray-green ones of the other island, but they grow so closely as to be almost in contact one with another. Some that project over the brink of the cliff look particularly bushy. Late in the day, hundreds of large petrels (Prerodroma neglecta), of which we have seen only a single example near San Felix, are circling about above the greenish plateau of San Ambrosio. The great bulk of them are of the dark phase, which is the opposite of the condition in the colony of the same species which we have visited at Ducie Island, in southern Polynesia. Flocks of the ternlets (Procelsterna) are flying about the base of the cliffs, the birds frequently alighting on pro- jecting rocks, close to examples of the booby, and also sometimes resting on the water in groups. Shortly after sunset, storm petrels (Fregetta grallaria) begin to appear over the ocean, fluttering about until darkness conceals them. Dr. Chapin’s observations and collections indicate that Pterodroma cookit defilippiana inhabits only San Felix and its outliers, while the larger petrel, P. neglecta, is confined to San Ambrosio. This is interesting because of a somewhat analogous division of adjacent island territories at Juan Fernandez by four species of the genus Prerodroma. In the latter instance, however, it so happens that P. neglecta and P. cookii defilippiana share the same island (Mas Atierra), while Mas Afuera is given up to another pair of species, likewise large and small, namely, P. externa and P. leucoptera masafuerae. (1) POLYNESIAN SOURCES OF WEST COAST WANDERERS The Chilean islets of Sala y Gomez, in latitude 26° 25’ S., longitude 105° 25’ W., are outside the South American region and the scope of this account. They may be mentioned, nevertheless, as a breeding station of certain pan-tropical PACIFIC SUB-TROPICAL ISLANDS 263 types of sea birds which have been unable to establish themselves in an exten- sive area of Humboldt Current influence. To find these species as breeding birds anywhere within about a thousand kilometers of the west coast of South America we must, in fact, go north of the equator. MacFarlane (1887, 209) visited Sala y Gomez on March 5, 1884. He states that the islets resemble St. Paul Rocks in the Atlantic, being a nearly inaccessible cluster of low black pinnacles against which the swell breaks high. The three species of birds he found in residence were noddies (Anoms), fairy terns (Gygis), and man-o'-war birds (Fregata). Easter Island, beyond longitude 109° W., has a limited but typical Polynesian avifauna. It is of present interest only as a possible source of the errant Sooty Terns (Sterna fuscata), elsewhere recorded from the neighborhood of Juan Fernandez and from the Peruvian coast at Independencia Bay. As we now know, however, these might have come from San Felix. 11. THe Desert Coast—Coquimso To Point PariNas To northward of Valparaiso the isotherms of both atmosphere and coastal ocean space out, further enhancing the equability characterizing the seaboard farther southward. Between Cape Horn and Valparaiso the mean air temper- ature increases approximately one degree Centigrade for each two degrees of latitude; but between Valparaiso and Callao the increase is less than one degree Centigrade for each five degrees of latitude. In fact, Antofagasta (24° S.) and Callao (12° S.) show a difference of only 0.7° C. in their annual average temper- atures. Compared with stations at the respective latitudes on the Atlantic side of the continent, the differences are even more striking than those of Patagonian South America. Thus the mean at Antofagasta is 4° C. lower than at Rio de Janeiro; that at Callao 5.6° C. lower than at Bahia. The last forests supported by rains appear on the coast along both banks of the Rio Limari. They are dense clumps of tique, already noted at Mocha Island, but here forming only smaller trees, well covered with epiphytes and with an undergrowth of ferns and mosses. Woods of this type fill up the gorges through the coastal plateaus but the remainder of the country is covered chiefly with brush and cactus. Most of the subsidiary streams along the lower course of the Limari, incidentally, enter through its left bank; on the right or northern bank is the waterless Quebrada Seca. As far northward as latitude 26° S., a winter rainfall system rules, but in progressively decreasing amount. The valley of the Guasco has no entirely dry years, but at Copiapé rainless years are frequent, even though occasional downpours between June and August give an annual mean of 21 millimeters. The winter mists, so characteristic of the Peruvian coast, also occur in Chile as far south as latitude 29° S. Known as gartas in Peru, and camachancas in Chile, their distribution evidently depends upon relief, for they are particularly marked south of Antofagasta where the slope of the coastal plateau is abrupt. According to Ball (1887, 129), they are practically unknown northward toward Iquique, which is the part of the coast attaining the maximum aridity. 264 OCEANIC BIRDS OF SOUTH AMERICA Between the Rio Copiapé, south of Caldera, and the Rios Camarones and Lluta, near Arica, there is a stretch of nearly a thousand kilometers in which only one river, the Loa, cuts the coastal pampa to flow throughout the year into the Pacific. It would be difficult to imagine a region more desolate than this portion of the Atacama Desert—a wind-swept surface, sunburnt and sterile, broken here and there by the grimy white of the “‘salares."’ Through it twines the hidden green canyon of the Rio Loa (Rudolph, 1927, 553). The true desert zone, interrupted only by the valleys of Andean streams, may be regarded as extending from the neighborhood of Caldera (27° S.) to Point Parifias (4° 30’S.). The lower limit of the summer rains on the western slopes of the mountains comes at about 2200 meters near the northern end of Chile and at 1600 meters in the latitude of Lima, while at Tumbez, in northern- most Peru, it reaches the sea. Throughout much of this zone, however, the desert is bordered toward the ocean by an ephemeral fringe of herbage composed of annuals brought into bloom for a few weeks by the winter mists. The islands on the coastal shelf partake, of course, of the climatic character of the mainland, but only those high enough to cause marked adiabatic cooling of the sea winds have sufficient moisture to share the transient vegetation of the ‘tiempo de lomas."’ The Chilean coastal islands are mostly inconsequential. The largest of those along the desert shore lie between Coquimbo and Caldera, plantless, waterless, and of potential value only as sources of guano. As to their occupation by the requisite number of sea fowl, I have no recent information. The island of Alacran off Arica has, however, been well covered with guano birds during the last few years. And what of the adjacent ocean along the arid shores of Peru and Chile? Darwin (1933, 329), when bound northward in the ‘Beagle,’ wrote: Here in the Pacifick, although the water is never agitated by storms, it never rests quiet, but feels through the unbroken continuity the violence which reigns in the South. Now, in the winter, a heavy dull bank of clouds intercepts during successive days even a glimpse of the sun. The temperature is by no means warm; in approaching these low latitudes I did not experience that delicious mildness, which is known for a few days in the Spring of England, or in first entering the Tropics in the Atlantic. A century ago Swainson (1836, 1, 261) could write, with a temerity which time has not justified, that “Chile and Peru are too arid and naked to excite great expectations regarding their ornithology."’ As a matter of fact, the bird life of the desert zone has proved extraordinarily rich and interesting. Aside from the land birds, moreover, the probability has developed within recent years that certain little-known sea fowl, including especially petrels and shearwaters of the Humboldt Current region, have their breeding grounds in remote parts of these continental wastes, where predatory animals are reduced toa minimum. Stresemann (1924, 61; 1929, 80) has called attention to mummies of downy young and adults of Hornby’s Petrel (Oceanodroma hornbyi) and of the Sooty Shearwater (Pxffinus griseus) found in the Chilean saltpeter fields and neighboring parts of the hinterland. No other nesting grounds of Hornby’s Petrel are known, and there are related species peculiar to the littoral ocean off THE ‘DESERT sCQAST 265 the west coast of South America, such as Oceanites gracilis gracilis and Oceano- droma markhami, the nesting stations of which are still undiscovered. The supposed nesting localities just referred to are (1), near Santa Luisa, 50 kilometers back from the coast and 1600 meters above sea level; (2), at a similar altitude 30 kilometers from the coast in latitude 22° 22’ S., longitude 70° W.; and (3), on the Pampa del Toco, inland from Tocopilla and north of the sites previously mentioned. Wetzel (1925, 284) reports many petrel mummies un- earthed through blasting away of the surface crust, and others in the bottom of a dry water-course. He assumes that at a period when the climate was some- what less arid, sea bird breeding grounds were widely distributed in the nitrate region, especially in the flat and relatively soft beds of the seasonally dry arroyos. Probably the nestlings, and even the adults of burrowing species, were surprised by sudden floods and buried beneath the rubble. Such floods still occur in the system of the Rio Loa. However, the evidence is not all related to the discovery of ancient birds. Sea fowl of sorts not yet properly identified still make their home in the dry valleys and gulches leading toward the nitrate pampas. An old resident of the Pampa del Toco informed Wetzel of a kind of *‘gull,’’ locally called **Gar- ruma,’’ which nests in such situations. The adults return only at night, bring- ing fish, and their cries are heard all through the hours of darkness. Such birds are unquestionably petrels. Furthermore, an engineer, Mr. Harry D. Ball, has written me from Tocopilla that on the “‘biologically bankrupt’ pampa, high above the narrowly green banks of the Rio Loa, there are “‘several nesting colonies of gulls or guano birds’’ upon which the wild dogs, or so-called foxes, of the Chilean desert prey at certain seasons. Mr. Ball continues as follows, with reference to his researches in this field: Frequently during the excavations and mining operations here I have come across “‘pockets”’ of guano. These pockets are loosely cemented sands, dark brown in colour and smelling very strongly of iodine. In these pockets can always be found small bird bones (chiefly the leg bones), although I have found one almost complete skull, bits of egg shell and very numerous brown insect carapaces. The guano-pockets are often found below 2 or 3 feet of hard cemented caliches and costras, and it is evident that they were formed at least previous to the final concentration and redeposition of the valuable salts of the pampa. These deposits of guano may throw light on the economically important problem of the origin of the nitrate deposits. Newton (1890, 375) long ago reported upon bones of small sea birds found in guano-like earth beneath the nitrate beds. A comparison of the limb bones, fragments of the bills, and other elements, suggested that they represented small petrels of the genus Oceanodroma, perhaps O. markhami. Philippi’s (1895, 11) original discovery of the present-day breeding grounds of Hornby’s Petrel, inland from Taltal (25° 30’ S.), was long obscured by the fact that he gave his specimens a new and hence unfamiliar scientific name. In a paper immediately following his description, the same author discussed (p. 14) bird bones from Mejillones, which had been found in guano of great age. For reasons which must surely be regarded as inadequate, Philippi believed the deposits to be much older than the underlying guano of the Chincha Islands in 266 OCEANIC BIRDS OF SOUTH AMERICA Peru, and perhaps to date from Tertiary times. He described five brain-cases of a booby as a new and extinct species, Sula antiqua. His plate, however, reveals no more than extremely slight or negligible distinctions from S. variegata, the common booby along this coast today. In nitrogenous guano from Tarapaca he found remains of a cormorant distinct from the four Chilean species with the bones of which he compared it. The supposedly extinct bird he named Phalacrocorax sulcatus, but the measurements, description, and figures indicate that it is nothing else than the modern P. bougainvilliiz. Philippi evidently misunderstood Taczanowski’s measurement of the length of the bill in the common Guanay, or Peruvian Cormorant, for the figures refer not to the culmen but to the distance between gape and tip. Another erroneous record for the desert region is that of a “‘penguin mum- my,’’ assumed to be several thousand years of age, and found under deep beds of guano at Huanillos, Chile (Lambrecht, 1933, 239). This specimen has even been assigned to a definite genus, Aptenodytes, but reference to the original source (Ribera, 1904, 339) shows that it was merely the downy chick of some sea bird, possibly a cormorant, and certainly not a penguin of any sort. The conclusions to be drawn from this discussion are that the ancient remains of sea fowl from the north Chilean desert region belong, so far as known, to present-day species of this coast; and, furthermore, that many facts of the utmost interest, including discovery of the first known nests, eggs, and young stages of a number of species, are likely to reward the first thorough reconnaissance within the area. As summarized by Aikman (1892, 439), one of the theories put forward to account for the great beds of nitrate of soda in the desert of Tarapaca is that the substance owes its origin to immense deposits of the guano of sea birds originally covering the shores of large salt lagoons. Such lakes; through an eventual overflowing of the shores, are supposed to have effected a mixture of guano with the salts and thus, by a process of slow decomposition, led to the formation of nitrates. The actual occurrence of small quantities of guano in the nitrate fields would seem to lend further support to this hypothesis. A serious objection to the theory, however, is the absence of calcium phos- phate from the deposits. It would be hard to account for the lack of such an insoluble salt while the easily soluble nitrate of soda remained. Again, if the theory were correct, we should expect to find further evidence in the form of portions of guano in the transition stage. Finally, there seems to be little doubt that the traces of birds’ nests, remains of birds, and the undifferentiated guano now found in the nitrate beds were left there subsequent to the formation of the sodium nitrate. Ball, however, in a manuscript communication quoted above presents evidence not wholly in accord with the last opinion. The most probable supposition, concludes Aikman, is one first advanced by Nollner to the effect that the nitric acid is to be ascribed to the decay of great masses of seaweed driven into the ancient lagoons at a time antedating the uplift of the land. The presence of large quantities of iodine in the raw nitrate of soda furnishes strong confirmation of this theory. Furthermore, the presence THE DESERT COAST 267 of small bits of still undecomposed seaweed within the salt constitutes addi- tional support. Smith (1904, 1) and Wetzel (1925) likewise discount the function of guano in the production of the still puzzling nitrate wealth. The former authority writes: Recent geologic changes are evidenced by the remains of 12 successive beach marks in 2,500 feet elevation on the hills back of the port of Mejillones. The amount of salts remaining is much too great for a simple impounding of an arm of the sea. But it is suggested that a tide may, at frequent intervals, have re-filled it, and that great accumulations of seaweed furnished the nitric acid necessary to convert the sodium into nitrate of soda. The presence of iodine (a component of seaweed) helps bear out this theory, as does the presence of fragments of undecayed seaweed. The location of the richest deposits on the sides of hills is further evidence of beach deposits. Briiggen (1934, 236) proposes a more advanced form of the earlier hypothesis, and holds that the nitrate deposits may be due to the action of bacteria, during an earlier period, in the presence of ammonia gas given off by abundant guano beds which formerly existed on the pampas. From Tarapaca to the northern end of Peru the coastland and the littoral waters are, on the whole, more uniform than along any equivalent stretch of shore line in the world. The relatively low temperatures for the respective latitudes continue northward from the tropic of Capricorn to Cape Blanco and the Galapagos. Antofagasta and Paita are farther apart than Cape Cod and Cuba, or than Norway and Gibraltar, and yet these Chilean and Peruvian ports are characterized by amazingly similar temperatures of air and of surface waters, which are affected little more by season than by latitude. This is the true zone of the Humboldt Current, of continuous upwelling, of secular uniformity in the desert shores, in the physical character of the littoral ocean and its associated life. As hinted previously, longitude here replaces latitude in divid- ing the climatic and faunal zones. I mean that one may skirt the land from Arica to Sechura Bay without meeting strikingly new diurnal or seasonal changes of weather and without finding more than trifling variation in the flora and fauna of the desert and the sea. An hour's journey to westward in the Pacific, on the other hand, will reveal changes more profound than a week's travel north or south. The Humboldt Current is responsible for a peculiar biota of its own and, from a map-maker’s point of view, zonal distribution may be said to become vertical instead of horizontal. So far as sea birds are concerned, the current establishes a magic wall between the cool-temperate inshore water and the outlying bands of oceanic water which are, successively, sub-tropical and tropical. The respective longitudes of the coast, of San Felix and San Ambrosio, and of Sala y Gomez have, as we have seen, their own distinctive and largely exclusive avifaunas. During the cruise of the ‘Vénus,’ a century ago, du Petit Thouars found ocean temperatures of 15.2° C. at Valparaiso, 16.7° at Callao, which is twenty- one degrees of latitude closer to the equator, and 17.7° at Paita, seven degrees of Jatitude still nearer the equator, during the early days of June. But in Post Office Bay at the Galapagos, at the end of June, the surface temperatures were 22.8° C. In other words, the temperature rose 5.1° C. during a journey covering 268 OCEANIC BIRDS OF SOUTH AMERICA four degrees of latitude and ten degrees of longitude between Paita and the Galapagos, although it had increased by only 2.5° C. throughout the previous voyage covering twenty-eight degrees of latitude and ten degrees of longitude between Valparaiso and Paita. From such observations de Tessan (1844, 443) concluded that in the Peruvian Current surface water is constantly being re- freshed along the length of the continental coast by water from lower levels. This was perhaps the first direct reference within the region to the phenomenon we now call upwelling. A comparison of thermometric conditions between places actually at the shore and those a short distance inland shows the effect of the ocean upon continental climate. It is especially marked during the verano or summer season. Thus at Caldera, Chile, on the coast, the mean annual temperature is 15.9° C., while at Copiapé (altitude 400 meters) it is 15.3°; but the mean for January at Caldera is only 19°, while that at Copiapé is 20.4°. The annual mean at Callao is 19.3°, at Lima (altitude 158 meters) 19°; but the February temperatures are, respectively, 21.2° and 23°. Santa Elena, Ecuador, has sum- mer means of from 23° to 24°; Guayaquil, inland on the river, 27°. The coast has a régime of regular alternating breezes, the ‘‘virazon’’ blowing toward the land during the day, and the ‘‘terral’’ toward the sea at night. The former wind is generally southerly, the terral an east wind. If one goes sufhi- ciently far offshore, the standard southeast trades will be found holding sway over the high sea. The most definite temperature effects produced by upwelling are conditioned by the topography of the submarine areas, for the sharpest lowering of temperatures of the surface waters corresponds in position with the oceanic deeps close to shore and with places where the continental shelf is reduced to a minimum. The morning gartas are produced by the abrupt cooling of air in contact with the sea at an hour when the land breeze still holds. Although these mists are particularly intense and seasonally prolonged in the south, they extend all the way northward along the coast to the Silla of Paita or, in midwinter, even to western Ecuador. In explanation of the normal inability of the air from the sea to bring rain to the land, Jefferson (1926, 285) writes that there cannot be much evaporation from the cool water or in the cool air that overlies it; if in the afternoon the land becomes hot enough to attract a sea breeze, the land must warm the cool air that blows against it enough to neutralize the adiabatic cooling produced by ascent to a consider- able height. That is, air at 65° F. may be warmed 25° by the land it is blown against if the ground has only the moderate temperature of 90°. It would take an ascent of nearly 5000 feet to cool the air by that 25°, leaving it nearly a mile in the air at the same temperature with which it started and with no progress made toward the condensation of water vapor. Almost the only onshore winds of the coast have this cause—the afternoon heating of the land; and the land is doubtless always hotter than 90° when the sea breeze sets in. The conditions are therefore very unfavorable to rainfall. But, although the rain forbears to descend except at intervals of decades, the “appearance of rain’’ is all but constant. Darwin in his ‘Beagle’ journal refers to the immense pall of still and low clouds covering the coast of Peru and northern Chile. Bowman (1916, 143) writes: THE DESERT