Bulletin of the Museum of Comparative Zoology AT HARVAED COLLEGE Vol. 123, No. 8 BIRD SPECIAT10N ON THE AUSTRALIAN CONTINENT By Allen Keast Australian Museum Sydney, Australia CAMBRIDGE, MASS., U.S.A. PRINTED FOR THE MUSEUM March, 1961 OCR text unavailable for this page.No. 8 — Bird Speciation on the Australian Continent By Allen Keast Australian Museum, Sydney, Australia CONTENTS I. Introduction 307 II. Materials and Methods 309 III. Definitions and Terminology 312 IV. The Australian Continent as a Biotic Environment 314 Physiography, Climate and Vegetation • ■ 317 Spatial Relationship of Australia to other Land Masses 323 V. The Australian Avifauna 324 VI. The Sampling of the Australian Avifauna 326 According to Classification and Family Categories . 326 According to Habitat (Vegetation formation) Occupied 327 According to Seasonal Movements 329 VII. Variation and Speciation in the Various Bird Families Podicipitidae 330 Accipitridae, Falconidae, Pandionidae 330 Ardeidae, Threskiornithidae, Ciconiidae 332 Anatidae 334 Rallidae 335 Turnicidae 336 Columbidae 337 Psittacidae 339 Coraciidae, Alcedinidae, Meropidae 347 Menuridae and Atrichornithidae 348 Campephagidae 349 Muscicapidae, Muscicapinae ■ ■ ■ ■ 350 Museicapidae, Timaliinae 356 Muscicapidae, Sylviinae 357 Muscicapidae, Malurinae 358 Artamidae 372 Sittidae 372 Dicaeidae 375 Meliphagidae 377 Zosteropidae 387 Ploceidae, Estrildinae 388 Oriolidae 389 Dicruridae, Cracticidae 390 Grallinidae 392 Ptilonorhynchidae 392 VI II. The Mechanism of Speciation on the Australian Continent 393 306 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY The Nature of Geographic Variation in Species. Quanti- tative Treatment 393 Examples of Species Formation and Intermediate Stages in the Speciation Process 394 (a) Superspecies 395 (b) Marginal Overlap of Recently Evolved Species . . 395 (c) Double Invasions 390 (d) Speciation by Circle Formation 398 Classificatory Position and Speciation 398 Habitat Specificity and Speciation 400 Seasonal Movements and Speciation 402 IX. Geographic Isolation 403 Terrestrial Isolating Barriers 403 (a) Rain Forest 40:; (b) Monsoon Forest 400 (c) Sclerophyll Forest 407 (d) Savannah Woodland 410 (e) Savannah Grassland 414 (f ) Mulga 414 (g) Mallee 415 (h) Desert (Spinifex) Grassland 415 (i) Stony (Gibber) Desert 415 (j ) Mangroves 416 (k) Swamps, Marshes and Streams 416 (1) Miscellaneous Habitats 417 (m) Bird Species ranging through many Vegetation Formations 417 Water as an Isolating Barrier 417 (a) Speciation between Australia and New Guinea 417 (b) Speciation between Australia and the Near Off- shore Islands 418 (c) Speciation between Australia and the More Re- mote Islands 419 (d) Radiation of Australian Species in the Archi- pelagos of the Southwest Pacific 420 X. Hybrid Zones and Stepped Clines — Evidences of Former Isolation 421 Hybrid Zones 421 Stepped Clines and Lesser Zones of Intergradation 423 XI. The Australian Environment of the Past and Former Speciation. Zoogeographic Subregions, the Refuge Con- cept, and Ornithological Evidence of Past Climatic Changes 424 Characteristics of the Tertiary, Pleistocene and Early Recent in Australia 424 Former Radiation and Speciation in Australia 426 The Zoogeographic Subregion Concept in Australia .... 429 KEAST: BIRD SPECIATION IN AUSTRALIA 307 The Refuge Concept in Australia 431 Harriers between Refuge Areas 435 Some Fragmental Avian Evidence supporting the postu- lated Quaternary Climatic Successions 436 Soil-Vegetation Relationships and the Climatic Past 438 XII. "Ecological" Speciation. The Development of Habitat Differences Between Species 43!) Geographic Variation in the Habitat occupied by Species 44(i (a) Habitat Versatility in Species with Broad or Generalized Ranges 440 (b) Habitat Variation in Species with Peripheral or Specialized Ranges 440 (c) The Occupation of Adjacent or Nearby Habitats by Different Infraspecific Forms 443 Ecological Differences between the Different Members of a Genus Inhabiting the same or Adjacent Habitats . . . 445 (a) The General Problem; the Situation in "Old" and Well-consolidated Species 445 (b) Recently-evolved Species. Habitat Differences in cases of Marginal Overlap and Double Invasion 445 (c) Interspecific Competition and its Influence on Habitat Plasticity 447 XIII. Continental and Archipelago Speciation Compared 449 Quantitative Differences between Continent and Archi- pelago 440 Qualitative Differences between Continent and Archi- pelago 450 Comparison of the Continent and Archipelago as a Physical and Genetic Environment 452 XIV. Summary 454 XV. Acknowledgments 457 XVI. References 491 I. INTRODUCTION This work is a study of vertebrate speciation on a continent. It seeks to understand the nature of the various forces operative in Australia, including the barriers breaking up distribution and leading to the isolation and divergence of populations. A quantitative approach is used to determine the number of new species currently evolving and the "potential" of the continent for producing further species. Our basic knowledge of the speciation process has been derived from the study of the faunas of archipelagos and the many striking illustrations to be found there provide the classic 308 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY examples. Continents, lacking the more bizarre intermediate forms and obvious distributional barriers (e.g. areas of sea), present a much less convincing case for geographic speciation. Thus, notwithstanding various fine studies, one school (Gold- schmidt, 1940, p. 183) categorically denies that the infraspecific variation that is such a feature of continental populations can be an intermediate step in the formation of new species. Such variation, it is maintained, is "dead end" and of a purely local adaptive nature. A large number of clear-cut examples of con- tinental speciation and the circumstances surrounding them are detailed here and explained. In order, moreover, to arrive at a quantitative comparison of the potential of archipelago and continent for giving rise to new species, the number of morpho- logically differentiated, isolated forms in Australia and an archipelago area of equivalent size are calculated. It is now generally accepted that the successive steps lead- ing to the completion of speciation are as follows: (a) Part of an individually variable population becomes spatially isolated from the parental stock. (b) Under conditions of isolation it diverges genetically, eco- logically, physiologically, and morphologically. (c) The differences become increasingly great. (d) Finally, the isolates become so different from each other that they can no longer successfully interbreed, i.e. they are reproductively isolated. The ultimate proof of this, of course, is only forthcoming in those cases when, under natural conditions, the two forms secondarily come together and do not interbreed. Mayr (1950a) has listed some of the proofs for geographical speciation as follows: (a) the occurrence of geographic varia- tion in species characters; (b) the existence of borderline cases between subspecies and species; and (c) "circular overlaps," that is, cases where the end members of a chain of interfertile races meet and do not interbreed. Recently completed speciation is indicated by examples of: (a) double invasions, (b) super- species, (c) marginal overlaps of closely similar species without interbreeding. Of particular significance also is the occurrence of allopatric hybridization, the result of a premature reuniting of differentiating isolates (i.e. before specific distinctness has been reached). The approach in the present instance is to investigate the occurrence of intermediate stages in the speciation process and their frequency, and to seek and detail cases of the important KEAST : BIRD SPECIATION IN AUSTRALIA 309 associated phenomena. Each species has been studied to deter- mine if it is monotypic or if geographic variation (clinal or in association with isolation) occurs. Isolates are graded accord- ing to the degree of their differentiation, and the total number for each species, family, and ecological subdivision calculated. The significance of various aspects of species ecology on varia- tion and speciation, in particular the vegetation formation (habitat) occupied and nature of seasonal movements, is treated in detail. Four hundred and twenty-five of the 531 breeding land and freshwater bird species and semi-species occurring in Australia (80 per cent of the total) are included in the present study. II. MATERIALS AND METHODS The present work was initiated in the period 1953-1955, when the writer was a Fulbright scholar at Harvard University. The basic material is from two sources: (a) systematic studies by the author (245 species) and, (b) recent taxonomic reviews, i.e. by workers using the "new systematics" (180 species). The bulk of the writer's work was carried out on the bird collections of the American Museum of Natural History, New York, which are easily the world's most comprehensive for the Australian and Pacific regions. They now include the Rothschild Collec- tions and the bulk of the Mathews types, as well as much recent material from Queensland (J. Henry) and Western Australia (K. Buller). Other U.S. collections examined were the Gould Collection in the Philadelphia Academy of Sciences, the Arnhem Land Collection (H. G. Deignan) in the U.S. National Museum, and recent Queensland collections (by K. Stager) in the Los Angeles County Museum. Following the writer 's return to Australia, material in the Australian Museum, Sydney (including recent collections made by the author in New South Wales, central and northern Australia), National Museum of Victoria, Melbourne, and South Australian Museum, Adelaide, was incorporated into the study. In the studies of geographic variation, standard taxonomic methods have been used (see author's various papers), in all cases the emphasis being on the population, not the individual. Published taxonomic data used in the work were obtained from the following sources: families Podicipitidae (Mayr 1943) ; 310 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Accipitridae, Falconidae, and Pandionidae (Amadou, 1941; Con- don, 1951a; Condon and Amadou, 1954) ; Ardeidae and Threskor- nithidae (Mayr, 1943 ; 1945b ; and Amadon, 1942 ; Amadon and Woolfenden, 1952) ; Anatidae (Mayr and Camras, 1938; Mayr, 1940a, 1945b ; Amadon, 1943 ; Amadon and Woolfenden, 1952 ; Ripley, 1942) ; Rallidae (Peters, 1934; Mayr, 1949) ; Psittacidae (Condon, 1941; Mayr, 1951; Cain, 1955); Sittidae (McGill, 1948; Mayr, 1950b); Cracticidae (Amadon, 1951); Grallinidae (Amadon, 1950a) ; Ptilonorhynchidae (Mayr and Jennings, 1952); and the following genera: Geopelia and Lophophaps (Mayr, 1951) in the Colnmbidae ; Acrocephalus (Mayr, 1948) in the Silviidae; Pachycephala (Mayr, 1954a); Falcunculus and Oreoica (Mayr, 1953a, b) ; Orthonyx (Mack, 1934a) ; Gerygonc (Meise, 1931); Malurus (Mack, 1934b; Serventy, 1951); Seri- cornis (Mayr, 1937; Mayr and Wolk, 1953) ; Acanthiza (Mack, 1936; Mayr and Serventy, 1938) in the Muscicapidae ; and Zos- terops lutea (Mees, 1957) in the Zosteropidae. In addition, the general study by Condon (1951b) on the birds of South Australia contains much basic taxonomic in- formation. Detailed studies of geographic variation were carried out by the author on the following : the families Turnicidae, Pediono- midae, Columbidae, Alcedinidae, Menuridae, Atrichornithidae, Campephagidae, Artamidae, Meliphagidae, Zosteropidae, Dicaei- dae, Oriolidae, the subfamilies Estrildinae (Ploceidae) and Muscicapinae (Muscicapidae) ; and the genera Megalurus (Silviidae), Climacteris (Sittidae), Orthonyx, Sphenostoma, Dry modes, Psophodcs (Cinclosomatini, Muscicapidae), and Epthianura, Ashbyia, Smicrornis, Aphelocephala, Hylacola, Calamanthus, Cthonicola, Origma, Pycnoptilus, Cinclorhamphus, Amytornis, Dasyornis, and Stipiturus (Malurini, Muscicapi- dae). The majority of these studies have now appeared, or will later appear, as separate papers. Those so far published are : Oriolidae and Megalurus (Keast 1956a, b) ; Aphelocephala , Dasyornis, Stipiturus, Alcedinidae, Climacteris (Keast 1957a-e) ; Muscicapinae, Amytornis, Epthianura and Ashbyia, Dicaeum, Artamidae, Estrildinae, Psophodcs, Smicroryiis, Campepbagidae (Keast 1958a-i). A marked advantage of using birds for studies in speciation is that there is a great amount of data on distribution, habitat and general ecology available in the literature. This applies equally well in Australia. Included in the standard works con- sulted in the present study are: North, 4 vols. (1901-1914), KEAST: BIRD SPECIATION IN AUSTRALIA 311 Mathews, 12 vols. (1910-1927), Serventy and Whittell (1951), Mayr (1941, 1944a, 1945a), and the periodicals, The Emu, and South Australian Ornithologist, now in their 58th and 44th year of publication, respectively. Field workers in various parts of Australia have kindly made supplementary material available. In addition, the writer has had the advantage of having worked on the birds of eastern Australia over a number of years and to have made field trips as follows: central Queensland (1950, 1951), Bass Strait islands (1951), Kimberleys, coastal Northern Territory, and north Queensland (1952), Victoria (1951, 1955), central Australia (1952, 1958), southeastern South Australia (1958), western New South "Wales, including the mallee regions (1956-1959), southwestern Australia and the Barlee Range (midwestern Australia) (1959). Besides acquiring taxonomie material, this has made it possible to check distributional limits, degree of attachment to habitat, aspects of seasonal movements, and to look at certain contact and hybrid zones. Information on the Australian continent as an environment lias been drawn from various official and other sources. In the case of the map showing former connections between Australia, New Guinea, and Tasmania, the 150 and 300 ft. isobaths are calculated from the sea depths shown on the Australian sheet of Carte Generate Bathymetrique des Oceans, 3rd Edition, 1942, Monaco. The map showing the limits of the Australian and Asian continent shelves is adopted from that in Mayr (1945c). Physiographic data are primarily from the 1942 Orographical Map of Australia by H. N. Warren, and Hills (1949). Data on temperature, rainfall, and associated factors, are from vari- ous official maps (undated), the Climatological Atlas issued by the Commonwealth Meteorologist (also undated), Leeper (1949), and the recently published Atlas of Australian Re- sources, Department of National Development, 1951-1955. The climatic map is based on the chart of Dr. J. Gentilli (undated). Other climatic maps, more precise and detailed, are to be found in Andrewartha and Birch (1954) and in the Atlas of Au- stralian Resources, cited above. The maps of vegetation forma- tions and soils are simplifications of those of Prescott (1931) and Wood (1949). Information as to Pleistocene and post- Pleistocene climates and vegetation changes is from Browne (1945), Crocker and Wood (1947), Gentilli (1949), David (1950), Fairbridge (1953), Crocker (1959a), and Wood (1959). 312 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY III. DEFINITIONS AND TERMINOLOGY For the purposes of the present study it has been necessary to depart somewhat from the terminology normally used by taxonomists. This applies particularly to the dropping of the term "subspecies" or "race" which, as used, simply means a morphologically differentiated population and tells nothing of evolutionary potential. The emphasis here is on isolation and hence the term "isolate" is used. Where subspecies names come into the text, it is as a ready means of referring to a form that alread}' bears a name in the literature. Species: The modern biological definition for the species is followed, the criterion being reproductive isolation. Care is taken to draw attention to, and view somewhat differently, those species that are members of superspecies (i.e. that are allopatric representatives), and that fall into the category of "semi- species" (see Mayr, 1942). The latter are the distinctive isolates of uncertain genetic status that the taxonomist may call species because they have reached a significant stage of morphological differentiation and for the sake of expediency. Isolate: This is a population or populations isolated from the main stock by a geographic barrier so that interbreeding is prevented. Isolates may be undifferentiated, slightly differenti- ated, moderately or well differentiated, the last-named being forms in which speciation is well advanced. In calculations of the number of morphologically differentiated isolates one stock is set aside as the "parent" from which the isolates are diverging. For the sake of consistency, in compari- sons between continent and archipelago, in those species that range through both, one form in each is set aside as "parental." Species that do not vary geographically, or have only clinal variation, do not, of course, have isolates. Population: This is used as an inexact term to mean simply the members of a species inhabiting a restricted area. It is credited with the morphological characteristics of a series of adult specimens from that area. Morphological characters: Those used are variation in colour- ation and colour-pattern, the dimensions of the appendages lowing length [a good eriterion of over-all size in birds], bill length, tail length) and, sometimes, the ratios of the length of one appendage relative to another. The term "morphological" is used to stress the fact that it is the variation visible in the preserved specimen that is being considered. KEAST: BIRD SPECIATION IN AUSTRALIA 313 Hybrid zones: These are characterized by the presence of very variable populations, with extreme types, as well as a range of genetic recombinants being present. Most are narrow but in some cases a rather broad zone of intergradation occurs. Hybrid zones mark the breakdown of a former barrier and the reunion of isolates that had reached a marked degree of mor- phological differentiation. A minor amelioration of climate in Australia, and in some cases human interference with habitat, is enabling various species to extend their range at the present time. Only a minority of them, however, have as yet contacted their geographic counterparts to form hybrid zones or zones of overlap. Stepped dines: The presence of a "step" in a cline is evidence of former isolation and there are good examples of this phenomenon in a few Australian bird species, e.g. Pachy- cephala pcctoralis (Mayr, 1954a). Minor zones of intergrada- tion that may not, or do not, indicate secondary junction (note also the remarks of Miller, 1955) are discussed separately. Continuous clines, though they may involve peripherally adapted forms that could diverge rapidly if isolated, do not form a stage in the speciation process. Hence they receive only passing attention. Barrier: This is an area within, or at the edge of, the range of a species where it cannot thrive because of its innate ecologi- cal limitations. The term is used here strictly as a geographic or spatial concept. The well defined distributional barriers, especially those that have given rise to distinctive isolates, are the critical ones here. Habitat: This term is used in a broad sense to mean the basic vegetation formation occupied by a species. About ten vegetation formations are of major importance to Australian birds: rain forest (under which may be included monsoon or pseudo-rain forest), sclerophyll forest, savannah woodland, savannah grassland, mulga, mallee, desert (spinifex) grassland, gibber desert, mangroves, and swamps and rivers. The majority of bird species occur in only one, or are common in only one, of these formations, sometimes being restricted to a specialized sector, association or sub-association within it (e.g. plateau-top heathland within the eastern sclerophyll sector). Habitat specificity is discussed in Section VI. The significance of the vegetation formation in bird distribu- tion and speciation lies, of course, in the fact that it represents 314 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY a zone of relative physical and biological constancy. It is to these grouped factors that species have adapted and specialised during their development. "Habitat" is as valid a species character as morphological or genetic criteria, in most cases. Refuge: This term refers to fertile or semi-fertile areas, mostly mountainous or hilly sections, in which plants and ani- mals have been able to persist during arid periods. Australia is currently in an arid phase, with marked aridity stemming from the early Recent. The rain forest, sclerophyll forest, and savannah woodland areas of the continent are today broken up into a disconnected series of peripheral segments, some of vast extent. Use of the term "refuge area" should not carry the inference that species "retreated" into them with the onset of aridity, but that only the stocks inhabiting those areas were able to survive. Sedentary or resident species: This refers to a species the members of which do not undertake seasonal movements but, as adults at any rate, spend the year in the general vicinity of their breeding territories. There is a minimum intermixture of indi- viduals from the different parts of the species range in this group. Migratory species: This term is used for species that have a south-north migration and regular breeding areas. They are typically inhabitants of the better watered coastal forests where conditions are uniformly good. Examples: Coracina novaehol- landiae (eastern Australia) and Artamus cyanopterus. Nomadic species: This term denotes those species that: (a) undertake extensive group movements of irregular amplitude and direction and, (b) breed wherever conditions happen to be suit- able, e.g. Epthianura tricolor and Histriophaps histrionica. (See also Section VI). Nomenclature: That of Mayr and Amadou (1951) is followed in the present work. The species order is that of the 1926 Check- list of the Royal Australian Ornithologists Union. IV. THE AUSTRALIAN CONTINENT AS A BIOT1C ENVIRONMENT. The Australian continent has a land area of 2,984,000 square miles, about three-quarters of that of Europe (including Eu- ropean Russia). It is approximately the same size as the United States and is some twenty -five times the land area of the British Isles. It has been a stable land mass for a considerable portion of its geological history and Archaean rocks outcrop over an KEAST : BIRD SPECIATION IN AUSTRALIA 315 extensive area. Australia differs from the other continents in various ways including : (a) Spatial isolation from other large land masses (Pigs. 1, - 1 — Wn ASIA CONTINENTAL SHELVES AUSTRALIA _M Fig. 1. Australia in relation to Asia, and showing limits of continental shelves (100 fathoms or 600 foot line). The clashes indicate Wallace's Line (original), the dots Weber's Line, and the continuous line the western limits of the Australo-Papuan continental fauna. At the height of Pleistocene glaciation the sea level only fell 300 feet. The cross-hatched areas nevertheless correspond fairly well with the extent of land at that time. (Figure modified from that in Mayr, 1945c). 2). (b) General flatness, mountain barriers to dispersal being virtually nonexistent (Fig. 3). 316 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY (e) Absence of areas of extreme cold, the winter snowbelt amounting to only a few hundred square miles. There are no permanent glaciers or snowfields. The over-all climate varies from subtropical to cold temperate. The mean annual temperatures Celebes .JESt -.^iik- »NEW GUINEA Timor ,'T.' ; a^s|H ^^^^ V. ,»tf New _-._.., r , •^H.br.des^flf'F'J' Won* SEA LEVELS LESS THAN 150 ff. m 300 -Pf ^fca^aNew Caledonia ,-j. Norfolk Is. '.»; Lord Howe Is. NEW ZEALAND '■^"'TASMANIA Fig. 2. Australia in relation to the islands to the north and east. The stippled areas indicate the land extensions that would follow a 150 foot drop in sea level, and the dashes show the 300 foot line. Tasmania and New Guinea were twice connected with Australia during the Pleistocene when the sea level fell 300 feet, Deep water isolates Timor, New Zealand, and the islands to the east and separates New Britain from New Guinea. range from over 80°F. in the tropical north down to 45°F. in the highlands of the south. Daily minimum temperatures for July (coldest month) are 60-70° and 30-45°F. respectively, and daily maximum temperatures for January (hottest month) arc 85-90° and 60-65°F., respectively. (d) Generalized aridity. Agriculturists classify Australia as being only one-third fertile, with the remaining thirds semi-arid and arid, respectively. Rainfall (Fig. 4) is the all-important factor governing the distribution of life in Australia. KKAST: BIRD SPECIATION IN AUSTRALIA 317 (e) The arrangement of the basic vegetation formations (which form the bird habitats) into broad parallel, north-south (mostly), or east-west zones (Fig. 5). The altitudinal zonation of plants and birds that is such a feature of the Rockies and other high mountain regions of the world is virtually absent in flat Australia. PHYSIOGRAPHY Pig. 3. Physiographic map of Australia. The continent is relatively flat. Mountains are not, per se, significant barriers to avian distribution. Their control of distribution, by way of rainfall and vegetation, however, is pro- found. The dotted line indicates drainage patterns. (Map redrawn from that of Mills, 1949). Physiography, Climate and Vegetation Physiography : The continent falls into three physical sub- divisions (Pig. 3) : the Great Plateau or Archaean Shield that covers the western half of the continent, the Central Basin, and 318 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY the Eastern Highlands (Hills, 1949; David, 1950). The back- bone of the last-named is the Great Divide, the range that extends southwards down the full length of the east coast, thence to Tasmania and westwards to central Victoria. It is by far the most significant physiographic feature on the Australian con- tinent and not only influences climate and vegetation over a wide RAINFALL AVERAGE Fig. 4. Rainfall of Australia (average annual figure in inches). The pattern is one of concentric, zones of increasing rainfall outwards from an arid central area. The desert extends to the coast in the south and west. The eastern, southeastern, southwestern sections, and a small area in the north, are fertile. (Map adapted and simplified from thai in the Atlas of Australian re- sources -Rainfall; Published by the Department of National Develop- ment, 1952.) area hut permits a richness and diversity in the flora and fauna that would otherwise be impossible. Nevertheless, only relatively KEAST: BIRD SI'KCIATION IN AUSTRALIA ;n:> restricted parts of it exceed 4000 feet in height and the highest peak in Australia, Mount Kosciusko in the southeastern section, is a mere 7,300 feet in height. Here and there over the surface of the continent, elevated areas, residual masses, and dissected plateaux take the form of mountain ranges. Amongst the most interesting of these from the faunistic viewpoint are the Mount Lofty and Flinders Ranges in South Australia, the Macdonnells and other ranges in central Australia, the Darling Scarp and Stirling Ranges in the south- Humid EZ3 Semi -arid Sub-humid (ZZI Arid CLIMATE iP? Fig. 4a. A simplified climatic map of Australia. Derived by J. Gentilli (Thornthwaite formula). west, Hamersley Ranges and the Kimberley district in the north- west of the continent (Fig. 3). Though typically these do not rise more than a couple of thousand feet above the surrounding plain thev have a marked local influence on rainfall and vegetation and 320 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY permit the existence of various forms of life that would otherwise he impossible. Rainfall: The rainfall patterns of Australia will be seen from the map of average rainfall (Fig. 4). The isohyets take the form of concentric zones of increasing rainfall extending outwards from an arid interior. The transition is particularly well denned in the east where the belts extend from north to south parallel with the coast. In places, tongues of dry country extend through from the interior to the coast, breaking up the more fertile country into sectors. VEGETATION H ( Mangrove Mangrove f 7 "?"^ Gibber desert | | Desert grassland (spin) 3 Mulga scrub |H 'l'l"l Savannah grassland Grass steppe [=1 Savannah woodland Mallee 15551 Sclerophyll forest [TTT] High moor |H Rain forest Fig. 5. Major vegetation formations of Australia, simplified from Pres- eott (1944) and Wood (1949). For explanatory notes see Table 1. There is a close relationship between the distribution of bird species and that of the major vegetation formations, with gaps in the formations acting as isolating barriers. M = mangroves; P = pseudo-rainforest (riverside). Two main systems operate to produce the Australian rainfall, a winter one that moves up over the southern part of the con- tinent (the "Antarctic System") and a summer one (the "Trop- ical System") that brings monsoonal rains to the north of the KEAST : BIRD SPECIATION IN AUSTRALIA 321 continent. The north of the continent has a wet summer and a dry winter, whilst in the south the reverse is the case. Since, however, the summer is hot in the north the rainfall there is less effective and the over-all climate is dry. The Eastern Highlands ensure that the east and southeast of the continent benefit both from the summer and winter systems and have a comparatively even rainfall. The central regions of the continent are towards the limits of penetration of both systems, either or both of which may fail. The 10-inch line in the south and 15-inch zone in the north outline the limits of the desert. Other facets of rainfall in Australia are its unreliability over large sections and the high rate of evaporation, factors greatly influencing plant growth. The irregularity of the rainfall in the interior explains the high proportion of nomads in the avi- fauna there. The climatic zones in Australia (Fig. 4a) reflect the rainfall picture. Vegetation: The basic vegetation zones occurring on the Au- stralian continent and a summary of their characteristics, are set out in Table 1. The approximate area of the continent covered by each (author's calculations) is given in Table 2. The close link between rainfall and the distribution of the basic vegetation formations in Australia will be seen from a comparison of Figures 4 and 5. Rain forest is restricted to the eastern seaboard and is broken up into pockets. Sclerophyll forest, eastern and southern in dis- tribution is also largely restricted to the coastal regions and is discontinuous. Savannah woodlands occur as broad tracts across the north and down the east of the continent (mainly). Savannah grasslands, mulga, and spinifex desert (desert grass- land) are the semi-arid to arid country formations. They occur as broad and continuous zones with the latter two occupying the central areas. Mallee, a stunted but floristically diverse formation, occurs in broad southeastern and western tracts (with some discontinuity) in the dry southern inland of the continent. Mangroves extend practically right around the Australian con- tinent but the rich and diverse tracts that support specialized mangrove bird species are restricted to the northernmost two- thirds. The swamps and marshes (not shown because of their 322 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY relatively small extent) and rivers (homes of the water-birds) are developed to by far the greatest extent in the better watered southeast of the continent (Victoria and New South Wales). Even so, a good proportion of them are temporary. ->=^- &£"^£$ ^_ 'B^s. JSP -£ftsL(- ■ r*^ ~£<*i-d ;H^? vfcoo oo l— "'.i„>*41jy . -OB" ■»«,>. ~~ ^'^» a • ■ o s iV, , j-O***^ a ■ • 8 e s £1 Grey and brown soils 1^3 Ston y deserts \Sij«H heavy texture 3 Tablelands and ranges hH Mallee soils and sandhills r— > Desert sandhills and ^777] Po dsols, red loams, etc. D <3 sandplains C^_- ItttJjj;! Low country, deltaic r /f/\ Red brown earths and terra rossas v'fiU'-' |:;;:; : ;!| Desert loams §V/\ Rendzinas and black earths |»«*»V] Brown soils of | High moor peat light texture Fig. 6. The soils of Australia. Map is a simplified version of that of Prescott (1944), reproduced by Taylor (1949). Definitions of soil types are given by R. Crocker (see Keast, 1959a). Physiography, rainfall, and soils, control the distribution of vegetation. Soils: The major soil types occurring in Australia will be seen in Figure 6. The characteristics of these have been summarized by Crocker (in Keast, 1959a). Attention to the fact that the zonal soil groups in Australia follow a pattern somewhat similar to rainfall has been directed by Wood (1959). Again, there is a broad, but not absolute, zonal sequence and correspondence be- tween soil type and vegetation. Soil history in Australia, and its influence on bird distribution, KKAST: BIRD SPECIATION IN AUSTRALIA 323 through the vegetation, will be discussed in Section XL Spatial Relationship of Australia to other Land Masses The distance between Australia and the nearest point of the Asiatic mainland today exceeds 2000 miles. This isolation dates back to the beginning of the Tertiary, a period of some 50,000,000 years. The limits of the Australian and Asian continental shelves (100 fathom or 600 foot line) will be seen on Figure 1. The former, it will be noted, extends out for only a short distance but the latter includes Borneo and Java, over one-third of the distance towards Australia. During the Pleistocene emergence the sea- level is believed to have twice fallen some 250-300 feet, and it is obvious that the Asian continent must have extended a consider- able distance towards Australia. Nonetheless, the faunistic dif- ferences between the two continents have been preserved. This underlies the significance of the intermediate deep-water gap, the islands within which have been highly unstable geologically. These islands, moreover, are "ecologically impoverished" and apparently only capable of supporting a somewhat limited num- ber of bird species. On these grounds also they represent a "dis- tributional bottleneck." A discussion of Wallace's Line is beyond the scope of the present paper. Several recent authors have, however, reviewed its zoogeographic significance ; these include Scrivenor et al.(1943), Mayr (1945c), and Darlington (1957). The islands adjacent to the Australian continent and their history are as follows : Timor: This lies some 350 miles to the northwest of Australia, a gap that was narrowed to perhaps 50-70 miles at the height of the Pleistocene emergence. As will be seen from Figure 1, how- ever, there is a narrow tract of deep water immediately to the east of Timor. The avifauna of Timor is a blend of species of western and eastern origin (Mayr, 1944a). New Guinea: This large island is only 100 miles from Australia at its closest point. The intermediate water is shallow, extending down to only about 30 feet in the section to the north of Cape York. New Guinea has repeatedly been in direct contact with Australia. This was the case during part of the Tertiary and on two separate occasions in the Pleistocene. New Caledonia: This island, 750 miles to the northeast of Australia, is separated by deep water and its isolation is of long standing. 324 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY New Zealand and Lord Howe Island: These islands lie, re- spectively, 1200 and 350 miles to the east of Australia. Both are regarded by geologists as "continental remnants," though a sub- marine trough extending down to 12,000 feet isolates New Zealand today. The nature of any pre-Tertiary connections with Australia are, however, highly speculative. Tasmania: This island lies about 140 miles off southeastern Australia but with islands half-way across, and has a purely Australian fauna. Intermediate ocean depths only extend down to perhaps 150 feet and, like New Guinea, Tasmania was directly connected to Australia during part of the Tertiary and Pleis- tocene. The influence of the surrounding land masses on the Australian avifauna has obviously been relatively slight, a fact that makes the continent ideal for speciation studies. Not only is the con- tinent surrounded by areas of deep sea but, if Tasmania be in- cluded with Australia, only to the north has there been "inter- ference" from outside. New Guinea has, however, continuously contributed new bird species to Australia. V. THE AUSTRALIAN AVIFAUNA Australia is credited with 707 species of birds by the Official Checklist of the Royal Australasian Ornithologists Union (1926). The recent treatment of Mayr and Serventy (1944), however, rightly reduces the number to some 651 species. There are 37 species of breeding seabirds, 83 visitors, 531 land and freshwater birds (29 of which come into the category of semi-species), with 14 being restricted to adjacent islands. It is on this total of 531 species that the calculations in the present work are based. In accord with its being the smallest land mass, Australia has less bird species than the other continents, as the following figures of Stresemann (1927-1934) will show: Australia 531 species Africa 1,750 South America 2,500 North America 750 Palearctic region 1,100 (Europe and Asia) The tropical island of New Guinea, immediately to the north of Australia, with only one-tenth of the land area, but dominated by tropical rain forest and high mountains, has no fewer than 509 species of land and freshwater birds, together with 28 semi- KEAST: BIRD SPECIATION IN AUSTRALIA 325 species, and 29 inhabiting adjacent islands, a total of 566 species. Thus it has a richer avifauna than Australia. If the 191 species common to both be eliminated, the combined total for Australia and New Guinea is 906, a sum that brings the section more into line with the major land masses of the Northern Hemisphere. The Australo-Papuan region is one of the major zoogeographic subdivisions of the world and, in conformity with its peculiar mammalian, amphibian, and freshwater fish fauna, is rich in endemic bird families, subfamilies, and genera. In contrast with some lower vertebrate groups (freshwater tortoises, hylid frogs), various invertebrates and some plants, and despite the occurrence of flightless ratites, numerous parrots, etc., on both, Australia has no close affinities with the avifauna of South America. Like- wise, supposed direct relationships with certain African elements (e.g., between Promerops and the Australian Meliphagidae) are now open to question (Mayr and Amadon, 1951). Many bird families do, however, extend from Asia to Australia, and some from Africa or Europe, through Asia, to Australia. Such may be equally well developed in both regions, e.g., Turnicidae, Campephagidae, Ploceidae, Zosteropidae. More commonly, how- ever, only one or two species of such Palaearctic or Oriental families extend to, or occur in, Australia : Meropidae, Pittidae, Timaliinae, Sylviinae, Motacillidae, Nectariniidae, Coraciidae, Sturnidae, Oriolidae, Dicruridae. Avifaunal interchange across Wallace's Line and down the "impoverished" Indonesian island archipelago has been discussed by Mayr (1945c), Darlington (1957), and others. The families and subfamilies endemic to the Australo-Papuan region have been listed by Darlington (1957) and Keast (1959b). Those occurring in Australia proper, and the number of species comprising each family, are as follows: Casuariidae (casso- waries), 1; Dromaeidae (emus), 1; Megapodiidae (incubator- birds), 3 ; Pedionominae (collared hemipodes), 1 ; Loriinae (honey lories), 6; Kakatoeinae (cockatoos), 10; Menuridae (lyrebirds), 2; Atrichornithidae (scrub -birds), 2; Malurinae (Australian warblers) , 64 ; Meliphagidae (honeyeaters) , 65 ; Cracticidae (Aus- tralian butcher-birds), 14; Grallinidae (mud-nest builders), 3; Ptilonorhynchidae (bower-birds), 8; Paradisaeidae, (birds of paradise), 3. Two hundred and thirteen species of Australian land and freshwater birds, or 35 per cent of the total, belong to these endemic groups. If, however, Australia and New Guinea be grouped together, with their total of 906 different species, the 326 BULLETIN : MUSEUM OP COMPARATIVE ZOOLOGY number of species belonging to these groups is found to reach 369 (41 per cent). In addition to endemic families and subfamilies Australia has various endemic genera belonging to cosmopolitan families such as the Anatidae (ducks), Columbidae (pigeons), and Ploceidae (finches). The Strigidae (owls), Cuculidae (cuckoos), Rallidae (rails), and others are each represented by various endemic species. In contrast to this, interestingly enough, a couple of cosmopolitan bird species extend to Australia, e.g., the owl Tyto alba, the falcon Falco peregrinus, and the grebe Podiceps crista- tus. Considerable radiation has occurred in the Australo-Papuan region in various cosmopolitan families. Thus, Australia now has 18 species of ducks and geese, 23 pigeons and doves, 51 parrots (though a proportion of these are lories and cockatoos), 10 king- fishers, and 12 cuckoos. From the viewpoint of its origin, the Australian avifauna could be said to have been built up by a series of waves of colonization from the north over a prolonged period, or by successive coloni- zations superimposed upon the earliest elements. Isolation has been fairly complete throughout. VI. THE SAMPLING OP THE AUSTRAL! AX- AVIFAUNA Eighty per cent (425 out of 531 species) of breeding land and freshwater birds are covered. The sampling of species in the present work, has been a random one. Only a couple of families of lesser importance are unrepre- sented. The proportion of species of New Guinea-Pacific and Australian origin in the sample approximates to that actually oc- curring, as does the proportion of species with the different kinds of distribution in Australia, e.g., southern as against northern, peripheral and inland, extensive and restricted. The ratio of species inhabiting the various basic vegetation formations and of those falling into the various categories of seasonal movement is likewise maintained in the sample. In the following, stress is placed on the relationship between speciation and (a) the phylogenetic group (i.e. family) to which a species belongs, (b) the vegetation formation it occupies, (c) the nature of its seasonal movements. Sampling According to Classification and Family Categories The classification and subdivisions of the Australia)! land and KEAST : BIRD SPECIATION IN AUSTRALIA .^27 freshwater bird fauna are set out in Table 4. The number of species in each family and sample used in the present study are shown in the first column. Thirty-one families are covered in toto whilst some 90 per cent of the Muscicapidae and more than half the parrot species are included. The only important families unrepresented are the Strigidae (8 species), and Cuculidae (11 species). Sampling According to Habitat (Vegetation Formation) Occupied The major vegetation formations occurring in Australia, and the approximate continental area covered by each plus its per- centage of the whole, are set out in Table 2. In succeeding columns the number of bird species restricted to or reaching its greatest abundance in each formation are shown, together with its percentage of the total avifauna. The final column represents a calculation of the number of square miles of habitat per species, to give a figure for the relative richness of each vegetation formation. The sample of each type used as the basis for all the calcula- tions in this work is set out in the first column of Table 5. Savannah woodland, savannah grassland, mulga, and spinifex desert, cover the largest areas of the continent. By contrast, rain forest and mangroves, though covering only a small area, are richest, on the one hand, in terms of number of bird species and. on the other, in number of species per square mile of habitat. The largest number of species, however, could be said to be specialized for life in sclerophyll forest and savannah woodland. This figure, it should be noted, does not represent the number of bird species that could be seen in a day's observation in a typical area of each habitat, for in a savannah woodland, partic- ularly, there is a relatively greater geographic replacement of species as between the north (tropical savannah) and the south. Rather, it expresses the number of species that are, basically, inhabitants of each. In the compiling of Table 2, it was apparent that the avifauna of the more specialized vegetation formations, the rain forest, scle- rophyll forest, mallee, desert (spinifex) grassland, mangroves, and swamps and rivers, tended to be fairly specific and conserva- tive in habitat choice. Hence, little difficulty was had in allocating species to a category. This did not apply, however, to the more 328 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY "transitional" types of vegetation formation: savannah wood- land, savannah grassland, and mulga, which not only have much in common but interdigitate and overlap extensively. The dom- inant species in savannah grassland (or savannah woodland) may also be common in savannah woodland (or vice-versa) and X X X X XX o o o o o o M. FasClOqulari S (mangroves) M. leujini (rain forest) M. chrijSOpS (sclerophyll) M. fuSCa (sav. ujoodlar^d) M. VireSCCnS (sov. grassland i. mulga ^ M. orna+a (maliee) Fig. 7. Distribution of (5 species of the genus Meliphaga (honey eaters) in eastern Australia to show correlation with major vegetation formations. The map is somewhat diagrammatic. Six species are inhabitants respectively of mangroves, rain forest, sclerophyll forest, savannah woodland, savannah grassland — mulga, and maliee. This is an exceptionally good example of a phenomenon seen to at least some extent in most bird genera. KEAST: BIRD SPECIATION IN AUSTRALIA 329 mulga, and penetrate the dryer sclerophyll forests and mallee. In these cases the alternative remained of creating several addi- tional "habitat categories" or placing the species under the vegetation formation in which they were especially prominent. For the sake of simplification the latter course was chosen. The figures for savannah woodland and savannah grassland are placed in brackets in Table 2 ; they are probably relatively too high. Apart from the above, a number of species occupy a life zone that is a continuum between two or more formations, e.g. the "leafy canopy," mistletoe. Again, large hawks, crows, small aerial feeding species (e.g. swallows), and a few others, are little influenced by vegetation in their distribution. They have a range that is virtually continent-wide. For the sake of completeness these are included in whichever of the "intermediate-type" habitat categories they are especially prominent. Each bird species is included only once in the calculations. The whole problem of allocating species to habitat categories is much simplified in Australia, fortunately, by the basic vegeta- tion formations being arranged in broad, parallel, zones. Sampling According to Seasonal Movements The seasonal movements undertaken by Australian land and freshwater bird species are complicated and every stage in the transition from one to another of the extremes is present. Thus, amongst the south-north migrants are species in which the move- ment is restricted to the more southern populations, or to a pro- portion of the individuals of these populations. In others the movement is entire. Amongst the nomadic species are some that are sedentary for a year or more, moving with the onset of a drought, and others that are nomadic in part of the range only. In the more typical nomads, however, the movements are general, irregular, occur frequently, and are of considerable amplitude. Many have, seasonally, a south-north bias to the movements. For the purposes of the present work, species are broadly grouped into the three categories according to the nature of the seasonal movements over the bulk of the species range. 330 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY . Grouped in this way, the Australian land and freshwater birds are as follows : Table 7 In Fauna In Sample No. of Species % of Whole No. of Species % of Whole The small number of true migrants and the relatively great importance of the nomadic way of life on the Australian con- tinent, as compared to the Palearctic and Nearctic regions, will be noted. Nomadism is an adaptation to unreliable and uncertain rainfall, though, in Australia, a certain amount of this behaviour is associated with the blossoming of food trees. Migratory species, in the restricted south-north sense in which the term is used here, are, as noted, mainly inhabitants of the well-watered eastern section of the continent. VII. VARIATION AND SPECULATION IN THE VARIOUS BIRD FAMILIES Order PODICIPEDES Family PODICIPITIDAE (Grebes) There are three grebe species in the Australian region, only one of which is endemic. All have an extensive Australian range, are nomadic, and are swam]) and river forms, diving for sub- merged food. Speciation Where continental variation occurs it is of a clinal nature. vide Podiceps novaehollandiae (Mayr, 1943). Isolation and dif- ferentiation does, however, occur in this species beyond Australia. Order FALCONES Families ACCIPITRLDAE, FALCON1DAE, PANDIONIDAE (Hawks and Eagles) The Australian eagles and hawks number 24 species and con- stitute a most varied fauna. At the one extreme are small "hov- ering" kites and a kestrel, at the other the large eagle Aquila KEAST : BIRD SPECIATION IN AUSTRALIA 331 audax with wingspan of up to ten feet. The majority of the forms are endemics (4 genera, 15-16 of the species) that, have diverged to a greater or lessor extent from counterparts in other areas of the world. The various forms are specialized ecologically in various ways : as scavengers, rodent and insect feeders, hunters that catch birds in the air, and as fish-eaters. Many are nomads and their distri- bution depends on seasonal conditions and the whereabouts of prey. All the Australian hawks are well differentiated species. Speciation Fifteen of the Australian Falcones either do not vary geo- graphically, or have only insignificant clinal variation. This results largely from the mobility that goes with large size and the strongly developed nomadic tendencies of most species. Of those that do vary geographically, one species Falco brri- gora, does so to a marked degree. It is improbable, however, that any of its forms are true isolates today (Condon, 1951a). Distinctive southern and northern forms occur in four species with evidence in at least one (Accipiter fasciatus), and possibly others (A. cirrhocephalus, Falco longipennis, Pandion haliaetus), that this state of affairs has been built up by invasion of the continent from the north taking place in two waves. In Accipiter fasciatus these consecutive arrivals are now connected by a stepped cline (Condon and Amadon, 1954). Aquila andax and possibly Accipiter novaehollandiae, show some evidence of interruption to gene flow across Bass Strait. The northwestern stock of Aviceda subcristata and the south- western population of Falco peregrinus, each of which has differ- entiated to some extent, are apparently isolated. Although there is evidence in a few species of minor isolation and range thinning, there would not appear to be any species being formed in the Australian hawks today. In the island section to the east of Wallace's Line, by contrast, the Australian species are broken up into about 43 morphologically differentiated iso- lates. This is exclusive of the variation occurring to the west of the Line in cosmopolitan, or Afro-Asian, species such as Milvus migrans, Haliastur indus, Falco peregrinus, and Pandion haliaetus. 332 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Order GRESSORES Families ARDEIDAE, THRESKIORNITHIDAE, CICONIIDAE (Herons, Spoonbills, Ibises) The 19 species that make up the Australian members of this order may be divided up as follows : — swamp bitterns, 2 ; man- grove bitterns, 1 ; egrets, 3 ; night herons, 1 ; stream and swamp herons, 4 ; reef herons, 1 ; estuarine herons, 1 ; spoonbills, 2 ; ibises, 3 ; storks, 1. All of these extend beyond Australia and a few be- long to widespread superspecies. The majority of the forms are predominantly swamp feeders but the herons hunt in the open fields, fringes of streams, and tidal flats as well. One species is confined to mangroves and a second is exclusively a reef feeder. The Australian Gressores almost all have a wide range on the continent and typically are nomads, moving around with seasonal conditions. Breeding areas vary somewhat from year to year, depending on water levels on individual marshes and streams (H. J. Frith, personal communication). A couple of species, however, are seaside feeders, and exhibit little seasonal movement. These are the Mangrove Heron, Butorides striata, and the Reef Heron, Demigretta sacra. Speciation None of the members of this group vary geographically except (a) Threskiornis mohicca, that apparently shows a slight south- north size cline, and (b) Butorides striatus, that has no fewer than five differentiated isolates (Mayr, 1943). Under circum- stances of insular isolation in the southwest Pacific, by contrast, Nycticorax caledonicus has 2 distinctive isolates, Dupetor flavi- collis at least 2, and Butorides striatus 4. The Australian isolates of B. striatus occur as follows: (a) Shark Bay (unnamed form), (b) Ashburton River to Point Cloates (rogersi), (c) King Sound (cinereus), (d) Northern Territory and Melville Island (stagnatilis), (e) Cooktown area {littleri), and (f) New South Wales (macrorhynchus) . These forms differ markedly in colour and size and one of them (rogersi) was long considered a distinct species. Mayr stresses the fact that the western and eastern groups are distinct, and suggests that they may have originated as a double invasion. The pattern of isolation and differentiation in B. striatus, set out in Figure 8, is similar to that seen in some other mangrove KEAST: BIRD SPECIATION IN AUSTRALIA 333 birds (see Section IX). Demigretta sacra has a similar distribution to Butorides stri- atus but does not vary geographically in Australia. It is a BUTORIDES STRIATUS (heron) 4 MALURUS CYANEUS - SPLENDENS GROUP (wrens) Fig. 8. Upper: Isolation and differentiation in the mangrove bittern (Butorides striatus). There is a chain of distinctive isolates along the western, northern, and eastern coasts (indicated by spotted areas and numbers). Eange gaps correspond to areas where the appropriate habitat is apparently lacking. Lower: Distribution of species and forms in the Malurus cyaneus — melanotics — splendens group of the blue wrens (see cross-hatched areas in lower half of map). The four major forms shown, though isolated from each other, are so distinct that (with the possible exception of M. melanotus and M. callainus) they must be regarded as (taxonomic) species. The habitats differ markedly, M. cyaneus inhabiting sclerophyll forest and thickets in savannah woodlands, M . melanotus, mallee, M. callainus, spinifex desert, and M. splendens, sclerophyll and dry scrubs. M. cyaneus has isolated insular races on the Bass Strait islands and Tasmania. "H" refers to a minor zone of intergradation. The small letters, "a " <:; monsoon forest, 1, major form; Tregellasia capita, rain forest, 1, minor; Pachycephala pectoralis, various habitats, 5-7, minor and major forms; P. inornata, mallee, 1, a minor form; P. lanioides, mangroves, 3, minor forms; P. simplex, mangroves, 1, major form; and Falcunculus frontatus, sclerophyll and savan- nah woodland, 2, major forms. In addition to the above there are several eases of isolation without differentiation in the Muscicapinae, e.g. Petroica chry- soptcra and P. multicolor in Tasmania. RHIPIDURA FULIGINOSA SCLEROPHYLL Fig. 14. Geographic variation in habitat preference in the long tailed flycatcher Bhipidura fvliginosa. This species, with essentially a peripheral range, occupies different habitats in various parts of the continent. This form of ecological variation indicates how, simultaneously with the develop- ment of genetic and morphological characters, isolates can become specialized for life in vegetation formations different from that occupied by the parental form. The numbers 1-9 on the map indicate races and isolates. The pairs of parallel lines represent distributional barriers (see Keast, 1958a). Note the isolate (9) in the mountains of central Australia. 354 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Bhipidura fuliginosa (Fig. 14) and Pachyccphala pectoralis provide interesting examples of geographic variation in the vegetation formation occupied, coincident with the acquisition of geographic morphological differences. In different parts of the range they inhabit rain forest, sclerophyll forest, savannah woodland, mangroves (and in the case of the latter even mallee) — see Section XII. Colonization by New Guinea rainforest species across Torres Strait has been the main way in which new flycatchers have been added to the Australian avifauna, the group being best developed in the tropics. Various stages of differentiation from parental New Guinea stocks occur in Cape York in the different species. Three recent colonizers, confined to the northern tip of the Peninsula, have yet to differentiate : Monarcha f rater, Microcca griseoceps and Tregellasia leucops. In Monarcha trivirgata, an immigrant New Guinea stock (albiventris) occupies the northern part of this peninsula and a well-differentiated Australian form (gouldi) the southern part. They possibly meet and hybridize. In Machacrirhynchus flaviventer, the descendents of an earlier wave of colonization are now isolated in the Cairns-Atherton rain forests (secundus) and the later ones on northern Cape York (flaviventer) . In the genus Arses there is a similar situa- tion but in this case the earlier form is now so distinctive that, though still isolated, it must be regarded as having reached species status (A. kaupi). The later invader, A. telescopthal- mus, has itself now differentiated from the parental New Guinea stock (race lorealis). There has been some reverse colonization of the savannah woodland areas of southern New Guinea by Australian species, the following having given rise to distinct isolates there: Bhipi- dura leucophrys (race melalcuca), and Microcca leucophaea (zimmeri) . Hybrid zones do not occur in the flycatchers except for some minor ones in Pachyccphala pectoralis (Mayr, 1954a). There are no nomadic species amongst the Australian fly- catchers. Four species, however, have migratory populations in the southeast part of their range. This partial migration has not prohibited the development of geographic variation in these species, two of them having morphologically differentiated isolates (Bhipidura rufifrons and Myiagra rubccula), and two KEAST: BIRD SPECIATION IN AUSTRALIA 355 varying clinally (Pachycephala nifiventris and Monarcha mel- anopsis) . Size and colour dines are developed in most sedentary Australian flycatchers that have, on the one hand, an extensive RHIPIDURA LEUCOPHRYS Fig. 15. The influence of isolation on south-north clines of decreasing size. Bergmann's Rule has a wide application in Australian birds. In the continuously ranging Rhipidura leucophrys southernmost and northernmost populations differ in size by 11 per cent. In Seisura inquieta, however, in which there is a gap in the range, the difference is 22 per cent (see Keast, 1958a). The letters a, b, and c indicate isolated populations of S. inquieta. 356 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY south-north range and, on the other, range through areas of widely differing rainfall. Thus, size clines occur in Rhipidura leucophrys, Petroica cucullata, and Myiagra rubecula. Seisura inquieta (Fig. 15) provides an interesting demonstration of the effects of isolation on a south-north size cline. Southern- most and northernmost populations in this species differ in size by about 22 per cent, the range gap being in northeastern Queensland. In the comparable Rhipidura leucophrys, in which the range is continuous, it amounts to only 11 per cent. Speciation and radiation in the islands of the southwest Pacific are taking place in three "Australian" flycatcher species (Table 5). These are: Rhipidura rujifrons (Mayr and Moyni- han, 1946), which has some 20 morphologically differentiated isolates, Petroica multicolor (Mayr, 1934), with about 12, and Pachyccphala pcctoralis (Galbraith, 1956) with over 50. The last-named is the richest of all bird species in number of races (Mayr, 1954a, p. 11). In addition to the above, the genus Petroica has given rise to many island forms in the New Zealand area (Fleming, 1950). Subfamily TIMAL1INAE (Babblers) The seven members of the Tribe Cinelosomatini covered in the present review are all inhabitants of the heavily forested regions except for the arid-country Sphenostoma cristatum and the mallee species Drymodes brunniopygia. Speciation Psophodcs is composed of two species, one (P. olivaceus) restricted to the rain forest and sclerophyll forest regions of the eastern seaboard and the other the ''ecologically-versatile" (P. nigrogularis) inhabitating the sclerophyll and mallee of the southwest, with an isolated relict race in the eastern mallee tract (Icucogaster). Keast (1958g, Figs. 1 and 2) has sug- gested a series of steps, associated with major climatic oscilla- tions, reflected also in Pachyccphala inornata-rufogularis, to explain speciation in the genus. These are: (a) a parental stock formerly ranging right along the southern seaboard; (b) isola- tion of the stock into eastern and western populations as the KEAST: BIRD SPECIATION IN AUSTRALIA 357 result of climatic deterioration, and/or edaphic changes; (c) ecological and morphological modification of the western popula- tion, the result of its being more exposed to the harsh environ- ment; (d) eastward colonization of the western mallee-adapted form, now specifically distinct (/'. nigrogularis) ; (e) isolation of P. nigrogularis into eastern and western forms as a result of the mallee becoming divided into two tracts. Psophodes is interesting in that it demonstrates a pathway of adaptation from life in the luxurious coastal forests to that iu the semi-arid mallee. There is evidence that the southwest has also given rise to dry-country forms in other bird groups. The genus Dry modes has a surprising distribution, one species inhabiting the southern mallee and the second (basically a New Guinea one) the rain forests of Cape York. The latter has, more- over, a minor isolate some 400 miles to the west on the Roper River (colcloughi) , the only bird species to have an outlyer confined to this section. The origin of this race, like the circum- stances of the original isolation of the species themselves, is obscure. Orthonyx is composed of two species, 0. temminckii and 0. spaldingii. The former has a surprising pattern of distribution, the rain forests of eastern Australia from the Illawarra dis- trict, New South Wales, to the Bunya Mountains, Queensland (with minor range gaps) and then reappearing, as a distinctive colour form, 1500 miles to the north in New Guinea. 0. spal- dingii is confined to the Cairns-Atherton rain forest tract. Tlu 1 monotoypic Sphcnostoma crisiatum shows only clinal vari- ation. Subfamily SILVI1NAE (Old World Warblers) The only Australian representatives of this Palaearctic sub- family (Mayr and Amadou, 1951) are Megalurus (2 species), Acroccphalus (1), and Cisticola (2). Acroccphalus and Cisti- cola are Ethiopio-Palearctic genera. Speciation The swamp-dwelling Megalurus gramineus has minor isolates in Tasmania and in the southwestern corner of Western Au- stralia, but M. timoriensis, despite an extensive peripheral range, does not vary geographically in Australia (Keast, 1956b). 358 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Acrocephalus arundinaceus has distinct western and eastern forms, whilst there is also possibly an isolate in the Kirnberleys (Mayr, 1948). Cisticola is represented by two species, both oi' which have a wide extra-Australian range and are only second- arily Australian. One of these, C. exilis, has some four colour forms within the continent and possibly fairly complete isola- tion as between the populations inhabiting the east and northwest of the ranges (Lynes, 1930; Keast, unpublished). The other, C. jmicidis, has isolated, differentiated populations in the Nor- manton and Darwin areas respectively. Subfamily MALURINAE (Australian Warblers) In contrast with the Muscicapinae almost all the members of the Malurinae occurring in Australia originated within the confines of the continent. Those genera that are restricted to Australia are Epthianura, with its monotypic derivative Ash- by ia (5 species), Acanthiza (10 species, plus one in the moun- tains of New Guinea), Amytornis (7 species), Malurus (7 species, plus one in New Guinea), Smicrornis (monotypic), Aphelocephala (3 species), Pyrrholaemus (monotypic) Hyla- cola (2 species), Calamanthus (2 species), Cthonicola (mono- typic), Origma (monotypic), Pycnoptilus (monotypic), Cinclor- hamphus (2 species), Daspornis (2 species), Stipiturus (2 species), Acanthomas and Oreoscopus, the last two monotypic derivatives of Scricornis. Gerygoiic (9 Australian species) and Sericomis (5 species) are equally developed in New Guinea and Australia. The island, for its part, has several endemic genera that are close relatives of those occurring in Australia (e.g., Todopsis). The Australian warblers are all small, basically insectivorous, species. They inhabit either the foliage or ground and low undergrowth, with most genera being specialized toward one or the other zone. Acanthiza and Scricomis, however, have repre- sentatives in both zones. Those that live in the undergrowth are mostly characterized by long upturned tails, e.g. Malurus, Amytornis, Stipiturus and, to a lesser extent, Hy la-cola. The Australian Malurinae total 62 species. KEAST: BIKI) SPECIATION IX AUSTRALIA 359 Speciation in the Major Genera A number of the Australian "sylviid" genera provide excel- lent demonstrations of speciation within the confines of the continent. They can best be considered individually. EPTHIANURA ALBIFRONS and ASHBYIA SUP. SP. .Fig. 16. Distinctive relict races of the chat, Epthianura eroeea (numbers 1-4 and spotted areas on map) are isolated in river valleys 800-900 miles apart. Distribution of this species, requiring dense ground cover and sub- marshy conditions, can only lie explained by the north of the continent formerly having been wetter and providing continuous, or near-continuous, habitat of the right type. The black ellipse indicates the range of Aslibyia lovensis, an inhabitant of the gibber deserts. Though fairly sedentary, its range is continuous and restricted. It does not vary geographically. Epthianura albifrons (see cross-hatched areas) is somewhat nomadic on the mainland (see A) and does not vary geographically. Interruption to gene flow by Bass Strait, however, has led to the development of a minor insular isolate on Tasmania (B). Epthianura and Ashbyia: This group, reviewed in detail else- where (Keast, 1958c), demonstrates the close link between the type of seasonal movements undertaken by species and their tendency to develop morphologically differentiated isolates. Of 360 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY the five species, Epthianura crocea is unique in being sedentary, widely ranging, and in having highly specialized habitat re- quirements. It has four isolated and well differentiated popula- tions in river valleys, respectively 600, 700, and 800 miles apart in the north of the continent. E. albifrons, somewhat nomadic and with a coastal and inland distribution varies only in the insular Tasmanian population. E. aurifrons, habitat general- ized, distribution interior (i.e., wide and continuous), nomadic but movements varying with area and seasonal conditions, has no geographic variation. E. tricolor has a generalized habitat and a continuous interior distribution. It is markedly nomadic and with a seasonal shift in abundance from south to north of continent. Breeding habits must accelerate gene flow. There is no regular breeding area but it breeds wherever conditions happen to be suitable. It may breed in autumn in centre and north of continent as well as in spring in south and it is a colonial nesting species. No geographic variation occurs. Ash- byia lovensis is restricted to arid gibber desert in centre of continent, a continuous but moderately restricted habitat. It is fairly sedentary. There is no geographic variation. The number of morphologically differentiated isolates in the Australian chats is thus four. Three of these are in Epthianura crocea (Fig. 16). Gerygone: Nine of the 15 species in this genus occur in Australia and they fall into perhaps 5 species groups. Three of these (G. palpebrosa, G. magnirostris, and G. chloronota) are of New Guinea origin. Within Australia the various species are adapted to habitats ranging from tropical rain forest to coastal woodlands, mangroves, and the arid interior. They are typically sedentary forms but two species, Gerygone olivacea and G. fusca, have (south-north) migratory races. Speciation is actively occurring in the genns. G. olivacea, a sclerophyll forest-savannah woodland species, has isolates in north Queensland (a minor form), the northwest (rogcrsi), and southern New (ininea (cinerascens) . The New Guinea G. hypoxantha, known only from Geelvink Bay, could be an early isolate of this species. The rain forest species G. palpe- brosa is broken up into three fairly distinctive stocks, inhabit- ing the tracts of this association on Cape York {palpebrosa) , Cairns-Atherton (johnstoni), and Mackay-Kockhampton {fla- vida). A second rain forest species, G. {igata) richmondi has KEAST: BIRD SPECIATION IN AUSTRALIA 361 isolated stocks (minor races) in eastern New South Wales- southern Queensland (richmondi) , the Bowen-Mackay area (amalia), and Cairns-Atherton area (mouki) . The mangrove species, G. laevigaster, inhabiting the northern coastline from Derby to Normanton, is broken up into three minor forms inhabiting, respectively, the mangrove tracts in the following areas: Derby-Napier Broome Bay (broomei), Port Essington- Melville Island (laevigaster), and the Roper River-Norman River (mastersi). G. cantator of eastern Australia, a form ap- proaching the degree of differentiation typical of a species, is apparently also a derivative of this stock. The distinctive infra- specific form tenebrosa, ranging from Carnarvon to King Sound, bears a similar relationship to the widely-ranging O. magnirostris (Meise, 1931). G. c. chlorunota of Arnhem Land is an Australian isolate of a New Guinea species. There are thus some 10 morphologically differentiated isolates in the genus Gerygone within continental Australia. A further interesting feature is that several of the "brown" species of Gerygone are relatively better differentiated on the basis of habitat and call notes than they are morphologically. This particularly applies to G. richmondi, G. fusea, and G. laevigaster. Insular forms of Gerygone occur in New Zealand and on Lord Howe Island (G. igata) . Smicromis: This monotypic genus lacks isolates but shows marked geographic colour variation, plus a minor south-north size eline. The occurrence of the colour forms in this species in broad belts from east to west across the continent shows an interesting correlation with rainfall and temperature. It has been suggested that specific climatic thresholds may operate to produce this rather interesting variation of the Gloger Effect (Keast, 1958h). Aphelocephala : The three species in this genus are, respec- tively, inhabitants of arid gibber desert (A. pectoralis and A. nigrocincta) and savannah woodland-grassland (.1. leucopsis). The gibber deserts, occurring mainly in two extensive south- north tracts, break up the savannah into eastern, central, and western sections, leading to minor isolates in the latter species. The others do not vary geographically. Acanthiza : The thornbills extend widely over the continent. As many as five species may co-exist in an area (e.g., about 362 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY Sydney) but are nevertheless well differentiated ecologically, particularly in feeding zone and/or the sub-association occupied. Acanthiza presents various interesting situations from the speciation viewpoint (Mayr and Serventy, 1938). There is a ease of speciation by double invasion, A. ewingi and A. pusilla diemenensis representing successive waves of colonization of Tasmania. Well differentiated isolates that could be said to be approaching species status include A. pusilla katherina and A. rcguloides squamata in northeastern Queensland. The total number of morphologically differentiated isolates in Acanthiza (Table 3) is about 15, most of them minor forms. In the sclerophyll and savannah woodland species isolates occur in : the Cairns-Atherton area (A. pusilla katherina, A. nana flava) ; central Queensland (A. rcguloides squamata) ; Tasmania (A. pusilla diemenensis, A. cwingi) ; Mount Lofty area of South Australia (A. regidoidcs australis, A. nana laetior, A. lineata clclandi). There is a form of A. pusilla confined to Kangaroo Island (zietzi), and the population of A. lineata there {chandlcri) agrees in colouration with that inhabiting Victoria, not the adjacent areas of South Australia. In addition to the above, A. chrysorrhoa lias a well differentiated isolate in the region of the Gulf of Carpentaria (normantoni) . The distinctive colour types of A. pusilla and A. chrysorrhoa inhabiting southwestern Australia are of uncertain origin. They could represent western outlyers of forest stocks from eastern Australia, or be derived by local selection (the area is one of high rainfall) from adjacent dry country stocks (Mayr and Serventy, 1938; Serventy, 1953). However, the southwestern type of A. pusilla lias secondarily extended right across the continent to New South Wales. In this section, accordingly, the coastal and interior races are quite distinct. The interesting relict isolates of the ground-dwelling Samphire Thornbill (A. iredahi) inhabiting, respectively, the arid in- terior (iredalci), heath country in eastern South Australia (hedleyi), and samphire adjacent to St. Vincent's Gulf (rosi- )tae), have been discussed by Condon (1954). Attention is drawn by Mayr and Serventy (1938) to two noteworthy characteristics of the species A. pusilla and A. chrysorrhoa. They are particularly versatile ecologically, habi- tats being occupied that range from rain forest fringes with an annual rainfall of over 100 inches per annum, to savannali woodland, mallee, and mulga, in the 10-inch rainfall zone. Each KEAST: BIRD SPECIATTON IN AUSTRALIA 363 is split up into a chain of wet, intermediate, and dry country raees. Sericomis: The rain forest-sclerophyll-savannah woodland superspecies 8. frontalis- is divisible into four basic forms which, in that they are largely isolated, may or may not be entitled to species status (Mayr, 1937). They inhabit the southwest (S. maculatus) and southeast {8. frontalis) of the continent, Tasmania (S. humilis), and Cape York — New Guinea (8. beccarii), respectively. 8. frontalis and 8. maculatus have given rise to a few hybrid individuals in South Australia (Mayr). S. maculatus, which occupies the dryer southwestern part of the continent, is broken up into 3-4 minor isolated forms, two of which occupy the Abrolhos and Recherche Island groups, re- spectively (Mayr and Wolk, 1953). The rain forest species 8. citreogularis and 8. magnirostris, ranging from New South Wales to the Cairns-Atherton region, each have a minor isolate at the northern end of the range. Hylacola: This genus contains two species so similar that they would be regarded as geographic races did they not occupy adjacent habitats in one restricted area. Of the two, II. pyrrhopygia is an inhabitant of heathy under-scrub within the sclerophyll formations of eastern and southeastern Australia. The other occupies this type of habitat in southwestern and South Australia, but in the eastern part of the range inhabits mallee. The contact zone, in the Bendigo area of Victoria, is where the two kinds of habitat come together. Calamanthus: This genus is broken up into two stoeks, a coastal one (C. fuliginosus) and an interior one (C. campcstris) , which differ significantly in size and colour. The two are largely isolated but intergrade in the Eyre Peninsula and Coorong sec- tion of South Australia, C. campestris moving down dry corri- dors into the coastal habitat of C. fuliginosus. The latter has minor isolates on Tasmania and in southwestern Australia. Amytornis: This distinctive genus constitutes one of the few true desert groups in the Australian avifauna, the various species being basically inhabitants of Triodia (porcupine grass). Their habit is secretive and they keep low down. Dense con- tinuous cover is 1 1 to prime requisite. Distribution tends to be broken up into "pockets." Dispersive capacities are poor. There are two species groups: A. tcxtilis and A. striatus. The genus is remarkable ( Keast, 1958b) for: (a) Bizarre changes in bill-form, from that typical of an insect-eater to the 364 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY heavy "seed-grinding," finch-like type occurring in the A. text- ilis group (Fig. 17). The degree of size and colour variation as between speeies in the A. striatus group (Fig. 18), is also ex- ceptional. The two species groups exhibit evolutionary changes A Icxlilis purnelli A. mock'slus A. goyderi AMYTORNIS TEXTILIS SUPER SPECIES Pig. 17. Differentiation in 1 > i 1 1 form coincident with speciation in desert grass wrens of the Amytornis textilis group. The three species show suc- cessive stages in the transition from an insectivorous to a granivoroug type of bill. reminiscent of those seen in the Galapagos and Hawaiian archi- pelagos, (b) There are no fewer than 4 isolated populations, so distinctive morphologically that they must be genetic species, yet each known from only a single locality: .1. (joijdcri (lower Macumba River, north of Lake Eyre), A. dorotheac (MacArthur River, Gulf of Carpentaria). A. woodwardi (Alligator River. Arnhem Land), A. housei (Charnley River, in the Kimberleys). Amytornis provides one example of secondary overlap by newly evolved speeies: i.e.. A. modestus and .1. textilis in the Maedonnell Ranges. The two obviously originated in the east and west of the continent, respectively. Geographic variation in the habitat occupied occurs in Amytor- nis striatus and .1. textilis (Figs. 17a and 18a). The former, a widely distributed "parental" 1 species occupying a range of KEAST: HIHI) SPECIATION IX AUSTRALIA 365 habitats, lias budded off a series of species around the periphery of its range that occupy rocky gorges only. In the A. textilis group there is the alternative situation of this species, occupying the western half of the continent, and the eastern A. modest us, each somewhat versatile in choice of habitat, separating out A. textilis •••A A. modestus A. goyderi SPINIFEX PLAIN SCRUB / AND THICKETS AMYTORNIS TEXTILIS SUPERSP OPEN PLAIN (SPINIFEX ■>) SCRUB AND TREES PORCUPINE GRASS IN MALLEE Fig. 17a. Amytornis textilis group of desert grass wrens. Distribution of species and their geographic variation in habitat is shown. The black square indicates the restricted range of the heavy-billed A. goyderi (now probably extinct.) A. textilis and A. modestus, relatively similar western and eastern counterparts, secondarily overlap in range (without interbreed- ing) in the mountains of central Australia. Each frequents a range of habitats except in the overlap zone, where one keeps to the rugged gorges and the other to the valleys. Here A. modestus is probably the secondary invader. The letters refer to minor racial forms of A. textilis, and the numbers to those of A. modestus. where their distributions meet. Thus, in the Macdonnell Ranges in central Australia, the former keeps to the rock surfaces of the gorges and plateaux and the latter to the areas of soft spinifex on the valley floor. 366 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY The isolated derivatives of A. striatus (A. dorothcac, A. woodwardi, and A. housci) could only have reached their pres- ent range at a time when desert (spinifex) replaced savannah grassland (an unsuitable habitat in that it lacks permanent cover). COLOUR KEY CHESTNUT OR RUFOUS INDEFINITE BROWNS Fig. 18. Members of the Aniytornis striatus superspecies to show the acquisition of striking size and colour differences. Though isolated, these forms have diverged morphologically to such an extent that there can be little doubt that they are genetic species. KEAST : BIRD SPECIATION IN AUSTRALIA 367 Dasyornis: The two species, D. broadbenti and D. brachyp- terus, whose habitat is dense coastal undergrowth, have a dis- continuous and relict distribution, each being isolated into two distinctive stocks, a southeastern and a southwestern one (see figure in Keast, 1957b). The southwestern forms are virtually extinct todav. ROCKY RIVER GORGES PORCUPINE GRASS ON SAND PLAINS PORCUPINE GRASS IN STONY GULLIES PORCUPINE GRASS ON STONY HILLS AMYTORNIS STRIATUS SUPERSR HIGH RANK GRASS IN TUFTS CLUMPS OF PORCUPINE GRASS IN MALLEE Fig. 18a. The range of the various forms in the A. striatum superspecies. The parental A. striatus (1, a-d) extends widely through the arid and semi-arid interior, whilst the distinctive derivatives A. dorotheae, A. icood- u-ardi, and A. housei (2, 3, 4) are isolated in the rocky gorges of the McArthur, Alligator, and Charnley Rivers, respectively. The habitats oc- cupied in different areas are shown on the outside of the map, vegetation formations that constitute distributional barriers on the inside. The three northern species could only have reached their present position at a time when desert spinifex grassland was continuous through to the northern seaboard. Stipiturus: A single superspecies is involved here, there being three major, isolated, stocks falling into two species. Their dis- tribution is shown in Figure 19. Stipiturus is an interesting genus in two ways: (a) because of the large number (5) of morphologically differentiated iso- lates in the species S. malachurus. These, occupying shrinking areas of specialized coastal habitat and cut off from the main stock in southeastern Australia, occur as follows : — Tasmania 368 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY (littleri), Kangaroo Island (halniaturinns) , Mount Compass area of South Australia (intermedins), southwestern Australia (westemensis) , and Dirk Hartog Island (hartogi). (b) It pro- vides an interesting demonstration of an ecological transition df S. m. malachuru s STIPITURUS MALACHURUS ai e + SUPER SPECIES Fig. 19. Isolation, speciation, and habitat differentiation, in emu-wrens of the genus Stipiturus. S. malachurus (1), a secretive species requiring dense undergrowth, is broken up distributionally into a series of minor isolates around the periphery of the continent (see a-f). Distributional barriers are areas of open country and sea. Distinctive isolated forms, approaching or that have reached species status (S. mal- achurus mallee — 2, and S. ruficeps — 3), occur in semi-arid and arid mallee and spinif ex desert areas, respectively. Eastern populations of S. malachurus are limited to sub-marshy areas, but in the southwest, where coastal undergrowth and spinifex lie in close proximity, adjacent populations demonstrate the complete ecological tran- sition to life in arid places. KEAST : BIRD SPECIATION IN AUSTRALIA 369 from life in well watered coastal areas to that in arid spinifex desert. The three forms occupy quite distinct habitats: 8. malachurus, subniarshy coastal heathlands and thickets (basically), 8. (mala- churus) malice, the mallee, and is) . The third species group is composed of two species, C. erythrops (inhabiting mountain sclerophyll in the east) and C. affinis (mulga desert). Climacteris provides one of the best demonstrations of dif- ferentiation and speciation in refuge areas. These same tracts of country also function as refuge areas in other groups, e.g. parrots, nuthatches. Ncositta provides a most interesting contrast with Climacteris in that (though differentiation has again occurred in these refuge areas) the various isolates have secondarily reunited to form hybrid zones (Mayr, 1950b). The genus has a chain of forms around the periphery of the continent, as follows : southeast {chrysoptcra chrysoptera) , southwest and south (pileata), north- west and north (Icucoptcra), northeast (striata), central east (albata, leucocephala, etc). Ncositta provides a splendid example of multiple hybridization (intergradation) along contact fronts (see Section X), and is surprising for the degree of differentia- tion achieved wivhout the attainment of reproductive isolation. The morphological characteristics, ranges, and approximate areas of hybrid zones in Ncositta are shown in Figure 21. Differentia- tion of forms has occurred in the same areas as in Climacteris, except that Ncositta lacks them in the Hamersley and Eyre Peninsula sectors and has several in eastern Queensland. Clinal variation occurs in various Australian Sittidae. Family DICAEIDAE (Flower-seekers and Pardalotes) This family, in Australia, falls into two genera, the endemic 376 BULLETIN: MUSEUM OP' COMPARATIVE ZOOLOGY and somewhat aberrant Pardalotus (7 species) and Dicaeum (1 species). The latter is well developed in southeastern Asia, Indo- nesia, New Guinea, and islands of the western Pacific. Dicaeum hirundinaceum, the Australian representative, is itself the end member of a superspecies extending through from Asia. Pardalotus occurs throughout Australia. P. punctatus inhabits the sclerophyll forests of eastern and southern Australia and has given rise to P. quadragintus in Tasmania and to the mallee species, P. xanthopygus. P. melanocephalus mainly inhabits the tropical savannah woodlands. P. rubricalus is a savannah grass- land-arid country species. The remaining species live in dry sclerophyll forest and savannah woodland, with a distinct ten- dency to occur along rivers. Dica( um hirundinaceum occurs throughout the continent and has an extremely wide habitat tolerance. Dicaeum is highly nomadic. Pardalotus striatus has a south- north migration. Other pardalotes. inhabiting the dryer areas, are apparently partial nomads. Speciation Pardalotus contains an instance of speciation by double inva- sion (into Tasmania), P. quadragintus representing the earlier, and the endemic race of P. punctatus (Icachi) the later one. Apart from this, the only differentiating isolates in the group are the insignificant southwestern form of P. punctatus and the northwestern one of P. melanocephalus. The puzzling status and relationships of Pardalotus substri- atus, P. or nut us, and P. striatus, largely geographically repre- sentative forms recognizable only on minor grounds, have been discussed by Ilindwood and Mayr (1946) and Serventy (1953). The first of these extends widely across the southern two-thirds of the continent, the second through a broad south-north zone in the east, and the third is confined to Tasmania and the south- eastern coastal strip of the mainland. The three overlap in eastern Australia but, despite their morphological similarity and the absence of habitat differences between them, only 1.8 per cent of the specimens are morphologically intermediate, i.e. hybrids, according to the findings of Ilindwood and Mayr. Serventy (1053), however, interprets the evidence differently, saying that ornatus is itself nothing but a hybrid form. He KEAST : BIRD SPl'.CIATION IN AUSTRALIA .!n also draws attention to the existence of a specimen from south- western Australia that has a yellow wing speculum (a charac- teristic of the southeastern striatus) as indicating the presence of occasional genes of that species in the far west. Serventy suggests that all the "striped-crowned" pardolates, hence, prob- ably belong to a single species. In any event, there can be little doubt that P. substriatus originated in southwestern Australia and /'. striatus in Tasmania. A high proportion of the latter occurring in eastern Australia are winter migrants. P. melanocephalus is the northern representative of the group. It obviously originated in the northwest or north and has sec- ondarily spread southwards to overlap the range of /'. sub- striatus. The ecology of Dicaeum hirundinaceum, its intimate relation- ship with the mistletoes (Loranthaecae), and the influence of these on its potential for developing geographic variation have been discussed by Keast (1958d). Loranthus berries form the main food of Dicaeum, and the bird undertakes extensive seasonal movements coincident with the fruiting of the plants. This may be in spring, summer, autumn, or winter, in different areas. Its breeding is also linked to the berry crop. The occurrence of mistletoes in all the main forest and scrub associations explains the wide range of vegetation formations inhabited by Dicaeum. The bird is a major disseminator of Loranthus. D. hirundinaceum has only negligible geographic variation in Australia. Where, however, the nomadic habit has been lost (i.e., in the colonizing of the Aru, Kei, and Tenimber Islands), distinctive insular isolates occur. Family MELIPHAGIDAE (Honeyeaters) This is an Australo-Papuan family of about 150 species, some 67 of which occur in Australia. Their common character is the brush -tongue and, though primarily insectivorous, nectar is prominent in the diet. Of the 20 or more genera occurring in Australia, 12 obviously originated here: Melithrcptus, Plectorhynchus, Certhionyx, Acanthorhynchus, Oliciphila, Eamsayornis, Granticlla, Conopo- phila, Zanthomiza, Phylidonyris (with Meliornis), Manorina (with Myzantha), and Anthochaera-Acanthdgenys. At the species level all are apparently Australian with the exception of about 18 that are recent immigrants from New 378 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Guinea, or are obviously derived from such. The majority of the 23 species of the large genus Meliphaga occurring in Aus- tralia are endemic. Honeyeater species are specialized for life in the full range of vegetation formations and the degree of "habitat tolerance'' of any one is commonly limited. Many species are sedentary but ''blossom nomadism," of limited amplitude, is widespread. A few species, especially Grantiella picta, Certhionyx variegatus, Myzomelia nigra, are highly nomadic. Meliphaga chrysops and Melithreptus lunatus are partial south-north migrants in the southeastern part of their range. Speciation The Australian Meliphagidae contain examples of a wide range of speciation phenomena, including two instances of triple in- vasion, one of re-invasion, and several demonstrations of sec- ondary range overlap by recently evolved species. There are a considerable number of minor instances of isolation and differ- entiation. (a) Instances of Recently Completed Speciation and some Special Phenomena (i) Melithreptus lunatus and M. albogularis, a case of sec- ondary range overlap in peripheral sclerophyll-savannah wood- land species (Fig. 23). These two species are so similar that, if they did not co-exist over an extensive area, they would rank as no more than minor geographic races. Both are strictly peripheral in range, inhabiting sclerophyll forest and savannah woodland. M. lunatus is southern and eastern in distribution, M. albogularis northern and eastern. The overlap area extends from Cairns to the Richmond River, a distance of 1,000 miles (Fig. 23). Interest in this case lies in: (a) circumstances of the original isolation and hence speciation; (b) how it is possible for two such similar species to co-exist. There can be little doubt that M. albogularis developed in the Ivimberley-Arnhem Land sector of the northwest, that is to the west of the Gulf of Carpentaria. A recent eastward extension of range by various northwestern bird species and races has been noted (Keast, 1956a). M. albogularis, however, has followed this with a great range extension southwards through the coastal forests of eastern Queensland. KEAST : BIRD SPECIATION IN AUSTRALIA MELITHREPTUS LUNATUS GROUP 379 M . albogularis pale blue unai Fig. 23. Speciation in the Melithreptm lunatus group of savannah wood- land — sclerophyll forest honeyeaters. M. lunatus (1 — stippled areas) is related to the others as follows: M. albogularis (2 — see cross-hatching), barely distinguishable on morphological grounds, mainly inhabits the north of the continent but overlaps the range of M. lunatus in eastern Australia by about 1,000 miles. M. affinis (3 — range shown in black), is a distinctive Tasmanian isolate. M. Iwnatus chloropsis (4 — heavily spotted area), a long-billed isolate in the southwest, has reached a lesser degree of divergence. There can be no doubt that the sibling species M. albogularis originated in northwestern Australia (or possibly even New Guinea) and that its occurrence in eastern Australia is secondary. In at least part of the overlap area it occupies a different habitat to M. lunatus. 380 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY M. lunatus and M. albogularis occupy slightly different areas in the overlap zone. The former occurs only on the tops of the ranges and the latter only in the lowlands (Barnard and Bar- nard, 1925; Barnard, 1926). Several other southern bird species arc restricted to the higher country in the northern parts of their range, e.g., Meliphaga lewini and Dacelo novaeguineae. In col- onizing southwards, M. albogularis may accordingly have been able to move, partly unimpeded by competition, through the lowlands. The fleshy orbital ring is a vivid orange colour in the eastern race of M. lunatus. In other members of the genus, and in the southwestern race of this species, this area of the body is a drab white, greenish, or pale blue colour. The character could have an important role in helping to prevent hybridization with the superficially similar M. albogularis. 1 (ii) Meliphaga lewini, M. notata, and M. gracilis, a case of triple invasion by rain forest species from New Guinea (Fig. 24). These three species have the same colouration and colour pattern and obviously are derived from a common New Guinea stock. The chief differences are in size, M. lewini being the largest and M. gracilis the smallest, and in the relatively long bill of M. gracilis. M. lewini does not have a close counterpart in New Guinea, whereas M. notata shares a semi-species relation- ship with M. analoga there (Rand, 1936), and M. gracilis, in Australia, is only racially distinct from its parental form. M. lewini, the species with the southernmost range (Dande- nongs to the Cairns-Atherton area), has obviously had a fairly long history in Australia. By contrast, the bifurcation of analog a -not at a from gracilis must have taken place in New Guinea. M. notata and M. gracilis are now each broken up into two populations in Australia, inhabiting the Cape York and Cairns rain forests, respectively. Where the three species overlap, M. lewini occupies the high- lands and M. notata the lowlands (Barnard, 1926). The long- billed M. gracilis is also a lowland dweller. i Brown and Wilson (1956) might regard this as an example of a character being modified as the result of contact between two closely related species. Against this, however, is the argument that the red orbital ring is typical of M. lunatus throughout its eastern range, not just where it meets M. albogularis. There would seem to be no doubt that the latter is the invader. Actually there is relatively little evidence of modification to, or reinforcement of, a morphological character in zones of contact between closely related species in Australian birds — though see Meliphaga virescens-versicolor. KEAST : BIRD SPECIATION IN AUSTRALIA 381 © M. nolafa © M. lewini MELIPHAGA LEWINI GROUP Fig. 24. Triple invasion of Australia by similarly coloured New Guinea rain forest honeyeaters of the Meliphaga lewini — notata — gracilis group. The distinctive endemic species M. lewini (1) obviously represents the first wave. If. notata (2), the second arrival, ranks as a semi-species with M. analoga of New Guinea. M. gracilis (3), the most recent arrival, is only racially distinct from its New Guinea parent. M. lewini and M. notata occupy different sub-zones where they overlap, the former keeping to the highlands and the latter to the lowlands. M. lewini has a south-north cline of increasing bill length and decreasing; wing length. 382 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY (iii) Meliphaga virescens and M. versicolor, a case of specia- tion in New Guinea and reinvasion of the Australian continent into a new habitat (Fig. 25). Fig. 25. Meliphaga virescens superspecies, a case of speciation by "re- invasion." The parental species has a wide range through the arid interior of Australia. It only enters mangroves in the west where conditions are dry. The New Guinea derivative, M. versicolor, has re-invaded the continent (see black arrows) to colonize the unoccupied eastern mangroves. It does not go beyond the forest fringes adjacent to the mangroves and hence has a range parallel to, but not overlapping, that of M. virescens. The dotted arrows indicate the range of an earlier specialized derivative, .1/. fasciogularis, a species inhabiting the more southern mangroves of eastern Australia. The numbers indicate geographic races of M. versicolor. KEAST: BIRD SPECIATION IN AUSTRALIA 383 M. virescens, a widely ranging savannah grassland-mulga species, varies clinally in colour and size. M. r< v si color lwis three forms, one inhabiting northern New Guinea that is barely dis- tinguishable from typical virescens, one in the Milne Bay area that is intermediate, and one in southern New Guinea and Cape York that has accentuated colouring and is markedly distinct from virescens. Nevertheless, all would rank as one species were it not for the fact that M. virescens and M. versicolor occur to- gether, without interbreeding, on Cape York. M. versicolor obviously represents a branch of M . virescens that became isolated in New Guinea and there built up genetic and morphological differences from the parental stock. The habitat relationships of the two are interesting. M. virescens extends throughout the dry parts of the continent but does not penetrate into the wet eastern coastal strip. In the dryer west, however, it does occupy mangroves. On Cape York it is confined to the dryer savannahs of the western side. M. veriscolor, a coastal species in New Guinea, has invaded Australia into the offshore mangroves of eastern Cape York. The mangroves of southeastern Queensland are occupied by a third related species, M. fasciogularis. This is isolated from M. virescens by the wet coastal forests but overlaps M. versicolor in range without interbreeding. (iv) Philemon argenticeps and /'. novaeguineae, a case of triple invasion from New Guinea (Fig. 26). The friarbird Philemon {moluccensis) novaeguineae has a wide distribution through the islands to the north of Australia. It has invaded the Australian continent three times, the earliest wave giving rise to the savannah dwelling P. argenticeps, and subsequent ones to P. novaeguineae gordoni and P. n. yorki in the mangroves (mainly) of Arnhem Land and coastal savannah woodlands of Cape York, respectively. The degree of morpho- logical distinctness of the forms is only moderate and were it not for their co-existing they would only rank as races (Mayr, 1944a, p. 167). P. argenticeps, the earliest arrival, is now well adapted to the dry country and extends well south into it. (v) Dry country Meliphaga species of the cratitia-plumula- <>rnata-keartlancli-flavesce7is-fusca groups, examples of "refuge area" differentiation and speeiation. These species fall into two or three species groups, the members of each of which tend to occupy different segments of the fairly :J84 BULLETIN- : MUSEUM OF COMPARATIVE ZOOLOGY ^^^ novae qui neae «52» P. nov. gordoni PHILEMON^ "5^ Fig. 26. Speciation in tin' genus Philemon (Meliphagidae). /'. novae- guineae has invaded Australia from the north three times. The firsl coloniza- tion (now P. argenticeps), adapted to dry savannah forest, has reached specific distinctness, as shown by its non-hybridization with later arrivals. Subsequently, there have been parallel waves of colonization into the north- west and northeast, respectively (novaeguineat gordoni and n. yorM). The latter two keep to the mangroves and wetter coastal savanah. The lower map demonstrates a case of the reverse process, colonization of the savannah woodland areas of southern Xew Guinea by the Australian savannah species, Philemon citreogularis. The insular population has now reached race status. KEAST: BIRD SPECIATION IN AUSTRALIA MS.! nvy (mainly savannah woodland-grassland) parts of the con- tinent. Their ranges, however, may be marked by broad over laps today. Habitat differences are often present but ill-defined with equivalent plant associations being occupied by different species as between north and south, or east and west. Current distribution patterns indicate that species arose in the east (fusca), north (flavescens) , and south (cratitia, ornata of the continent. (vi) Myzantha flavigula and M. mclanotis in the Victorian mallee. This puzzling case has been discussed by Serventy (1953). In southwestern Australia the widely ranging, savannah wood- land-grassland M. flavigula is connected clinally with a darkly pigmented race in southwestern Australia (ooscura) , inhabiting the coastal high rainfall zone. In the Victorian mallee, however, where a similar colour form (melanotis) occurs, it behaves as a distinct breeding unit co-existing, but not hybridizing, with M. flavigula. The circumstances of the original isolation of M. mclanotis are obscure, unless the species is of southwestern origin. (vii) The semi-species Meliphaga melanops and M. cassidix, secondary contact without interbreeding. These two species are so similar morphologically that they can only doubtfully be distinguished in the field, il/. )itis, by contrast, are Australian savannah woodland species that have given rise to forms in southern New Guinea. KEAST : BIRD SPECIATION IN AUSTRALIA 387 Hybrid zones do not occur in the Meliphagidae, though the relationships of the northern and southern forms in Melithreptus gularis merit investigation. As noted, relatively few of the Australian Meliphagidae extend beyond the continent, but those that do tend to have morpholog- ically differentiated insular isolates. Thus, Myzomela dibapha has about eight, mostly in the Celebes-Timor sector but with one on New Caledonia. Lichmcra indistincta falls into many forms (over 5) in a range extending from Timor to the south- west Pacific. Other species have minor isolates in New Guinea. Clines are common in the Meliphagidae, with most being dem- onstrations either of the Bergmann or Gloger effects. Meliphaga lewini, however, has a south-north cline of increasing bill length. The "plastic" interior species, M. virescens, provides examples of somewhat complicated colour variation. In Melithreptus gularis and Meliphaga penicillata, as in Smicromis brevirostris, three colour forms extend in broad zones from east to west across the continent. As noted, they are associated with specific rainfall and temperature threshholds. Family ZOSTEROPIDAE (White-eyes) Of three Australian species of Zosterops, Zosterops lateralis is equally well developed on the continent and in the southwest Pacific, and the other two (Z. lutea and Z. chloris albiventris) are derived from island species to the northwest of the continent. Z. lateralis has an eastern and southern distribution in Aus- tralia and is essentially an inhabitant of sclerophyll forest and rain forest. Z. lutea, inhabiting the north and west, is a man- grove dweller. Z. chloris albiventris is confined to the islands off Cape York. The species are sedentary except that Z. lateralis has a limited south-north migration in the southeast of the continent. Speciation Z. lateralis has morphologically differentiated isolates in southwestern Australia (gouldi) and on the Capricorn Islands off eastern Queensland and Lord Howe Island (tephropleura) . The Tasmanian form (tasmanica), also occurring in southern Victoria, and that occupying the Murray Valley (halmaturina) , apparently former isolates, are now connected with nominate 388 BULLETIN: MUSEUM OF < O.M PAR ATI VE ZOOLOGY lateralis (eastern New South Wales) by minor zones of inter- gradation. The Cairns-Atherton form (vegeta) may also be an isolate. Zosterops lutia has forms in the Hamersley and Kim- berley segments that are probably isolates. There is clinal varia- tion in both. Z. lateralis tasmanica has invaded New Zealand and thence Norfolk Island, the Chathams, and Campbell Island, within recorded history. The species has some 6 morphologically dif- ferentiated isolates in the southwest Pacific. The successive invasions of Lord Howe Island and Norfolk Island by members of the genus Zosterops provide some of the classic cases of speciation by multiple invasion. Whether or not these forms, however, came from a single point source is doubtful. Family PLOCEIDAE Subfamily ESTRILDINAE (Finches) Of the 18 species composing the Australian Estrildinae only two, and possibly three, are derived from outside of the con- tinent. Lonchura flaviprymna, inhabiting a restricted area in the northwest, is an isolated derivative of L. maja of Indonesia, and Erythrura trichroa has recently colonized Cape York from New Guinea. Lonchura castaneothorax has a wide range in New Guinea (two morphologically differentiated isolates) as well as in Australia. The seed-eating finches are mainly inhabitants of the dryer creekside thickets and savannahs of the north and east. Zonae- ginthus bcllus-oculatus, by contrast, is confined to thickets in the sclerophylls of the south, Estrilela temporalis to the sclero- phyll and rain forests of the east, Lonchura castaneothorax to rcedbeds and areas of rank grass (east and north), Zonae - ginthus pictus to rocky outcrops in the central desert, and Erythrura trichroa to rain forest fringes in the northeast. Whilst none of the finches undertake regular seasonal move- ments, those inhabiting the central and northern regions are somewhat sensitive to shifts in the supply of seeds and surface water, bad seasons forcing groups to move to new areas. Speciation Zonaeginthus bi<>- luccanus), Alcedinidae {Halcyon sancta and H. chloris), Cam- pephagidae (Coracina tenuirostris), Muscicapinae (Arses kaupi, Monarcha melanopsis, Hcteromyias cincreifrons) , Malurinae (Gerygone (igata) richmondi), Dicaeidae (Dicacum hirundi- naceum), Meliphagidae (Myzomela dibapha, Mcliphaga notata) , and Ploceidae (Erythrura trichroa, Lonchura flaviprymna) . Many superspecies groups are confined to the Australian continent. These include those in the following families and genera: Columbidae (Geophaps), Psittacidae (Platycercus, Bar- nardius, Pscphotus, and Neophema) , Muscicapinae (Petroica), Malurinae (Gerygone, Sericornis, Amytomis, Stipiturus, and Malurus), Sittidae (Clima-cteris), Dicaeidae (Pardalotus), Meli- phagidae (Melithreptus, Acanthorhynchus, Mcliphaga, Antho- chacra), Ploceidae (Zonae ginthus, Poephila), and Cracticidae ( Gymnorhina, Strepera). (b) Marginal overlap of Recently Evolved Species Some examples of this are as follows: — Psittacidae: Neophema elegans, N. chrysostoma, and N. chry- sogaster, overlap in parts of Victoria and South Australia. Areas of maximum concentration today suggest that N. elegans has arrived, secondarily, from the west, and N. chrysogaster from Tasmania. Campephagidae : Coracina robust as and C. papuensis co-exist 396 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY in the Cairns-Atherton area of Queensland, the latter presum- ably being the more recent arrival. Muscicapinae: Petroica rosea and P. rodinogaster now breed together in southern Victoria. The latter developed in Tas- mania. Pachyccphala rufogularis and /'. inornata overlap in the niallee of Victoria, the latter apparently being a secondary in- vader from the west of the continent (Fig. 13). Malurinae: Amytornis textilis and A. modestus both breed in the Macdonnell Ranges, the former having colonized the area from the west, and the latter from the east (Fig. 17a). Hylacola cauta (western) and H. pyrrhopygia (eastern) now overlap distributionally in the Bendigo area of Victoria. Sittidae: CUmacteris picumnis (melanota) and C. melanura co-exist at the head of the Gulf of Carpentaria, the latter being an invader from the west (Fig. 22). Dicaeidae: Pardalotus substriatus and P. melanocephalus live together in coastal eastern Queensland, the latter obviously being an invader from the north of the continent. Meliphagidae: Melithreptus lunatus and M. albogularis occupy the same spatial relationship to each other as do Pardalotus but in this case the overlap zone extends for several hundred miles (Fig. 23). Myzantha melanotis and M. fiavigula occupy adjacent zones in the Victorian niallee area, the former appar- ently being an "old" mallee form and the latter the recent invader. Cracticidae: Cracticus torquatus and C. mentalis co-exist in Cape York, the latter possibly being a recent colonizer from New Guinea. The co-existence of Menura novae-hollandiae and M. alberti in adjacent areas in northeastern New South Wales and south- ern Queensland is also an instance of range junction or overlap between related species though, in this case, obviously not a very recent one. (c) Double Invasions These', of course, are special cases of range overlap. They occur in two main areas: northern Australia, and Tasmania. There are also one or two apparent instances in southwestern Australia. Entry into the Australian continent is only possible at a couple of points : through Cape York and, to a lesser degree, KEAST: BIRD SPECIATION IN AUSTRALIA 397 through the northwest corner. Tropical invaders can mostly only colonize down the coastal rain forest and monsoon forest tracts. This enables a fairly clear picture to be gained of the significance of double invasion in the speciation of Australian birds. Colonization of northern Australia from the New Guinea- Timor segment: (i) Meliphaga lewini, M. notata, and M. gracilis (honey- eaters). These three rain forest species now co-exist in the Cairns-Atherton area and represent successive waves of, and degrees of differentiation from, the ancestral stocks in New Guinea (Fig. 24). (ii) Philemon argenticeps and P. novaeguineae (honeyeat- ers). In this case P. argenticeps, extending widely over the north of the continent, represents the first wave of colonization. Subsequent waves have independently entered Cape York (P. novaeguineae yorki) and Arnhem Land (P. novaeguineae gor- doni) (Fig. 26). (iii) Coraciva roousta and C. papuensis (cuckoo-shrikes) are examples of an "old" Australian species (but whose ancestors must have entered from the tropics) being secondarily contacted, in the north of its range, by a later invader. (iv) The two species of teal ducks (Anas eastanea and A. gib- berifrons), and the hovering kites (Elanus scriptus and E. notatus), represent somewhat older cases of speciation by double invasion. They now co-exist over a broad area. (v) Arses kaupi and A. telescophthalmus lorealis (flycatch- ers) and Meliphaga macleayi and il/. chrysotis (honey eaters). In these cases the earlier and later arrivals, though isolated from each other in the Cairns-Atherton and Cape York sections. respectively, are so distinct that they must be genetic species. (vi) Various infraspecific forms. "Simultaneous" invasion of Arnhem Land and Cape York by different races, which remain isolated in Australia, can be seen in Butorides striatus, Myzomela crythrocephala, and others. In a few cases a second wave of colonization by a stock, entering the continent at a common point, now hybridizes with the earlier one. In Accipiter fasciatus such hybridization occurs over a broad front. In a few species, the later invader has not yet contacted the earlier one. Monarcha trivirgata is an example of this, the New Guinea race occupying Cape York and the Australian one the country 398 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY to the south of the Princess Charlotte Bay dry barrier, as in "(v)" above. Colonization of Tasmania from Australia: (i) Acanthiza cwingi and A. pusilla (Malurinae). In this ease A. cwingi represents the derivatives of the earlier wave and A. pusilla (race diemencnsis) the later one. (ii) Pardalotus quadragintus and P. punctatus (Dicaeidae). This is a parallel case, with P. quadragintus the earlier element. (iii) The two insular "species" of Strepera (arguta and fuliginosa) could represent a double invasion of the island by one or the other of the mainland species, i.e. either S. versicolor or S. graculina. Most workers feel, however, that a representa- tive of each is involved. Colonization of the sclerophyll forests of the southwest corner of the continent from the southeast: Double invasion of the isolated pocket of sclerophyll forest in southwestern Australia by the robin genus Eopsaltria. would appear to be the explanation for the occurrence there of two species, E. georgiana (representing the descendents of the earlier wave) and E. australis (griseogularis) . Speciation in the Malurus lamberti complex could be accounted for by at least one of the species advancing from the east around the head of the Great Australian Big-lit. '.- ■ (d) Speciation by Circle Formation The only real instance of this in Australia is in the parrot Platycercus elegans (Fig. 11). The honeyeater, Mcliphaga versicolor, in that it developed its differences in New Guinea and from there has reinvaded Australia to occupy zones adjacent to its near relatives 31. virescens and M . fasciogularis, presents, to a degree, a parallel in the north of the continent (Fig. 25). The white-breasted races of the finch, Neochmia phaeton, have had a similar origin. Classificatory Position and Speciation As will be seen from Table 1, summarizing the results in terms of families, there is a considerable amount of variation in the "speciation potential" of the different bird groups. The hawks have only about 3 isolates in Australia and they are poorly differentiated ones. The larger water birds (grebes, herons and ducks) demonstrate similarly weak tendencies. In the case of the herons only 5 isolates are developed in Australia. KEAST : BIRD SPECIATION IN AUSTRALIA 399 The ducks have no continental isolates though several occur in the Rallidae. Large water birds have a very slow evolutionary rate, Howard (1946) having shown that many of them have not changed skeletally since the Miocene. The pigeons (Columbidae), inhabiting a wide range of habitat types, have 10-11 isolates in Australia and, since they total 22 species, approximately reflect the average for the avifauna as a whole (0.5 isolates per species). The parrots (Psittacidae), equally diverse in habitat requirements, have 17-18 isolates in 33 species. Many of the species in this family, are, however, geographically representative. The ratio of isolates to species in the kingfishers (Alcedinidae) is 4 to 10. In the cuckoo-shrikes (Campephagidae) there are 7 isolates in 8 species (1.0 per species). This is a figure typical of man}* of the small passerine groups, e.g. Malurinae with 61-62 isolates in 69 species, and the Muscicapinae with their 33-35 isolates in 43 species. The latter groups, it should be noted, are largely composed of sedentary species. The Meliphagidae have 18 isolates in 67 species, although here again there are a number of geographically representative species. The Estrildinae, most of which are restricted to the north, have 6 isolates in 18 species. The Artamidae, which have continuous ranges and are nomadic, have no isolates in 6 species. The endemic Australian Cracticidae, all of which are rela- tively large birds, have 11-12 isolates in 11 species. In this case, however, 4 of the latter are actually of uncertain genetic status so that the actual figure may be 1.5 isolates per species. The family to which a species belongs can be seen to influ- ence its potential for giving rise to new species not through any innate genetic characteristic so much as through what might be described as its "ecological attributes." Included in these are body size and general mobility (large birds require a larger foraging area than small ones), nature of seasonal movements, type of food ' ' niche ' ' occupied, and habitat. Small insectivorous birds commonly live in a habitat that is specialized and re- stricted, and have poor dispersive capabilities. The influence of the ■'history" of a group on speciation pat- terns is seen in the interesting differences between the flycatcher subfamilies Muscicapinae and Malurinae. The bulk of the genera in the latter are true endemics and speciation is taking 400 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY place almost wholly within the continent. The Muscicapinae, however, whilst they do contain endemic genera, are best de- veloped in the tropics and most of the isolation and speciation is taking' place either outside of Australia or between New Guinea and Australia. Habitat Specificity and Speciation As noted, the majority of Australian bird species are re- stricted to, or reach the peak of their abundance, in one or another of the basic vegetation formations. Thus, as will be seen from Table 2, 16 per cent are rain forest dwellers, 16 per cent live in sclerophyll forest, 2 per cent in mallee, 4.5 per cent in mulga, 3 per cent in desert (spinifex) grassland, 3 per cent in mangroves, 14 per cent in swamps and marshes, and so on. The remainder are less specific but nevertheless are oriented to one or the other of the •'intermediate-type" habitats of savan- nah woodland (28 per cent) and savannah grassland (11.5 per cent). That the habitat category into which a species falls has an important influence on its potential for developing isolates and hence undergoing speciation will be seen from Tables 5 and 6. Thus the 70 rain forest species sampled have 31 morphologically differentiated isolates (0.5 per species). In addition to some recently developed geographically representative species (in Tasmania and southwestern Australia) the 80 sclerophyll forest species have 71-72 isolates (1.0 per species). The 113 savannah woodland species have 61-64 isolates (0.5 per species), the 44 savannah grassland species 10 isolates (0.2 per species). Ten mallee species have 4 isolates (0.4 per species). Fifteen mulga species have no isolates. Fifteen desert grassland (spinifex) species have 8 isolates (0.5 per species), 16 mangrove species, 10-11 (0.6 per species), and 54 swamp species only 11-12 (0.2 per species). The 9 species placed in the miscellaneous habitat category (e.g., Lophophaps plumifera, Epthianura crocea) have 7-8 isolates (0.8 per species). If the figures be expressed as percentages (Table 6) it will be seen that of the total isolates about 14 per cent are rain forest forms, 33 per cent are sclerophyll forest birds, 30 per cent belong to savannah woodland, 5 per cent to savannah grassland, 2 per cent to mallee, nil to mulga, 4 per cent to desert grassland (spinifex), 5 per cent to mangroves, 5 per cent to swamps and marshes, and 3 per cent to miscellaneous habitats. KEAST : BIRD SPECIATION IN AUSTRALIA 401 One of the significant facts revealed by Table 6 is that the bulk of the morphologically differentiated isolates or "incipient species" in Australia are coming from species that are pre dominantly inhabitants of sclerophyll forest and savannah woodland. These account for 33 per cent and 30 per cent, re- spectively, a total of 63 per cent of all isolates. In contrast with the tropical rain forest and mangrove habitats (that are mostly restricted to the northern fringe of the continent) these are "true Australian" vegetation formations. They have the com- bination of being extensive, biotically rich, and yet are broken up into several isolated tracts. In these regards they contrast with the dry -country vegetation formations, the savannah grass- land, mulga, mallee, and spinifex, and with the swamp-river habitat. The highest proportion of isolates that can be described as well differentiated, as compared to slightly differentiated, occur in savannah woodland (32 per cent), sclerophyll forest (25 per cent), and rain forest (22 per cent), respectively. Hybrid zones are most numerous amongst the inhabitants of the savannah woodland, sclerophyll forest, and savannah grass- land habitats, accounting for 54 per cent, 24 per cent, and 18 per cent of the figures, respectively. The savannah areas, occu- pying "intermediate" climatic zones are presumably the most sensitive and responsive to minor climatic shifts. Sclerophyll forest and savannah woodland have, more than any other asso- ciations, been subject to alteration by human settlement. Species in which the only variation is of a clinal nature are most numerous in savannah woodland (43 per cent), sclerophyll forest (14 per cent), and savannah grassland (16 per cent). Species that do not vary geographically are also most numer- ous in savannah woodland ( many of the inhabitants of which are nomads, with others restricted either to the north or south), swamps (most inhabitants nomadic), rain forest (most are New Guinea immigrants restricted to the far north), and sclerophyll forest (many restricted to part of the southeast or Tasmania). Species with isolates beyond Australia belong mainly to four habitat categories in Australia : mangroves, rain forest, sclero- phyll forest and savannah woodland. In the case of the first two of these, this is to be expected as these vegetation formations have a wide distribution through the tropics. That so many sclerophyll and savannah woodland birds are involved might seem strange. The explanation is that the species concerned 402 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY are those that already exhibit a fairly wide degree of habitat and ecological versatility in Australia, so that adjustment to a new habitat in the islands apparently presents no problem. Examples, and the number of isolates beyond Australia, are : the hawks (Aviceda snbcristata (12), Accipitcr novaehollandiae (19), A. fasciatus (6) ; the cuckoo-shrikes Coracina papuensis (9) and C. tenuirostris (18) ; the whistler Pachycephala pcctor- alis (51) ; the wood-swallow Artamus lencorhyn chits (7) ; the dove Geopelia striata (3) ; the honeyeater Philemon novaeguineae (8) ; and the white-eye Zost crops lateralis (6). The rain forest Rhipicliira rufifrons, by contrast, is an example of a species with a restricted habitat tolerance in Australia but that exhibits ecological versatility in the southwest Pacific where a number of diverse islands are occupied. Seasonal Movements and Speciation The influence of seasonal movements on isolation and specia- tion is summarized in Table 7. Two-thirds of the sedentary and migratory species vary geographically. Only one-quarter of the nomads, however, do so. Almost all of the isolates, as well as the hybrid zones, occur in the sedentary species. The number of morphologically differentiated isolates and the number of isolates per species for the three categories is as follows : Number of Number of Morphologically differentiated isolates isolates in Australia per species Sedentary species 294 (69%) 191-196 0.65 South-north migrants 32 (8%) 15-17 0.50 Nomads 99 (23%) 7 0.07 It will be seen that the nomadic way of life virtually elim- inates any chance of isolation and speciation developing. This is brought out in equally striking fashion by the individual bird groups. The hawks and large water-birds are mostly nomads. Within the parrot family the nomadic Melopsittacus (Fig. 9) and Leptolaimus do not vary, whereas most genera composed of species that are essentially sedentary or resident, like Psopho- tus and Platycercus, have many isolates. Parallel cases are the nomadic Certhionyx and GrantieUa amongst the honeyeaters, as compared to most members of Mcliphaga. Epthianura, amongst KEAST: BIRD SPKCIATION IN AUSTRALIA 403 the Malurinae, and Artamus, in the Artamidae have, on the one hand, sedentary species that show geographic variation and, on the other, nomadic species that do not show any at all. South-north migration, by contrast, does not preclude the development of geographic variation and isolation, in part be- cause many of these undertake seasonal movements over part of the range only, e.g. Zosterops lateralis and Gerygonc olivacea. The relationship between variation and south-north migration in the campephagids Coracina novaehollancliae and C. tenuiro- stris has already been discussed. A reasonable inference with respect to most true south-north migrants in Australia is that, since they tend to inhabit the fertile coastal strip, they are able to return to the same areas to breed from year to year. This contrasts with the situation in nomads like Epthianura tricolor and Lalagc tricolor which, though there is a south-north com- ponent to the movements, concentrate for breeding where sea- sonal conditions are maximal. IX. GEOGRAPHIC ISOLATION An understanding of the barriers initiating speciation in Australian birds necessitates a detailed review of the distribu- tion of the Major vegetation formations in the light of those bird species living in them. Terrestrial Isolating Barriers (a) Rain Forest Rain forests are confined to the eastern seaboard and are broken up into a series of tracts and pockets (Fig. 5). There are, however, three main areas: (i) Cape York, (ii) Cairns- Atherton, (iii) and South Queensland-northern New South Wales. Each of these has some striking ornithological elements, the Oairns-Atherton section being particularly rich in this re- spect. (i) The Cape York rain forest tract: These forests are chiefly of significance in that many New Guinea species have recently colonized the northern section. There are some 15 of these species, most undifferentiated: Amaurornis olivaeeus, Probosciger aterrimus, Larius roratus, Geoffroyus geoffroyi, Syma torotoro, Podargus oceUatus, Cacomantis castanciventris, Drymodes superciliosus (also on Roper River), Microcca grisei- ceps, Trcgellasia Jeucops, Monarcha frater, Scricornis beccarii, 404 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Craspedophora magnified, Glycichaera fallax and Mcliphaga chrysotis. In several polytypic rain forest species the Cape York form represents a recent, independent wave of colonization from New Guinea, e.g. Megaloprepia magnified. The only species endemic to the Cape York rain forests is the honeyeater Trichodere cockerelli. There are, however, sev- eral distinctive races that have differentiated there, e.g. Arses telescophthdlm us loredlis. Ornithological data suggests that the Cape York rain forests have had the following history. They figured as "bridges" in earlier waves of colonization of the Cairns-Atherton forests by New Guinea species, but subsequently became reduced to such an extent that they could no longer support some of these forms. They have recently expanded again, 1 ut not to their former extent. The evidence for this is that a couple of genera, and species, requiring dense rain forest conditions, occur in New Guinea and in the Cairns-Atherton area, but not now in the intervening Cape York rain forests, e.g. the robin Hetero- myids, the log-runner Orthonyx and, apparently, the swiftlet CoUocalia spodiopygid. Again, the endemic bower-bird, Priono- durd newtoniana, has New Guinea affinities. It is presumed that these forms became secondarily exterminated from the Cape York area. Tate (1952) finds similar distributional anomalies in the mammals of northeastern Australia, and feels too that these can only be explained in the above manner. An interesting alternative suggestion is, however, advanced. This is that, when Torres Strait was dry land and the climate was wetter, the confluence of the rivers flowing southwards from Papua must have approached that of the Australian ones flowing northwards across the Gulf of Carpentaria. The accompanying riverside forests would conceivably have provided a migration route from New Guinea to Atherton that by-passed Cape York. As noted, there is currently much colonization of northern Cape York by New Guinea species suggesting re-establishment, or expansion, of this habitat. (ii) The Cairns-Atherton rain forests: These are quite ex- tensive and, in terms of the number of species endemic to them, are of considerable significance. There are some 8 such species • Heteromyias cinereifrons, Colluricincld boweri, Arses haupi, Oreoseopns gutturalis, Meliphaga frcnata. and M. ntaclcayana, Prionodura newtoniana and Sccnopoeetes dentirostris (both bower-birds). Some 25 rain forest and sclerophyll forest species KEAST : BIRD SPECIATION IN AUSTRALIA 405 have isolates of infraspecific rank here. They include Acanthor- hynchus tenuirostris, Ninox novae-scclandiae, Climacteris leuco- phaea, Eopsaltria capito, Pachycephalia pcctoralis, Gerygone (igata) richmondi, Acanthiza nana, Acanthi za pusilla, and Opopsitta diophthalma. There can be no doubt that the elevated and dissected Atherton area is a "refuge" of long- standing. The rainfall today exceeds SO inches per annum and in one small section readies 160 inches per annum. It is probable that even in the event of a marked reduction in overall continental rainfall this sector would still be covered with rain forest. (iii) The southern rain forest tracts: These extend from about the Bunya Mountains in southern Queensland to Dorrigo in central New South Wales. Thereafter they taper out as a series of small pockets. They are most dense and extensive in the Lamington area on the Queensland-New South Wales border. Only between one-third and one-half of the true rain forest species extend south to this last sector. Within the 400-500 miles of the tract, moreover, there is a steady falling off in the number of species. A few, however, extend right through to the restricted, cold, subtemperate rain forests in the Dandenongs in Victoria, e.g. Meliphaga lewini. Amongst the species endemic to the southern rain forests and associated wet sclerophylls are the bower-bird Sericulus chryso- cephalus, the scrub-bird Atriehornis rufescens and the frog- mouth Podargus (ocellatus) plumifcrus. A lyrebird species, Menura alberti, is confined to the elevated Lamington segment in the north, where there is also a race of the southern sclero- phyll whistler, Pachyccphala olivacea (macphersiona) . (iv) Barriers and Speciation: Speciation occurs mainly be- tween the three major tracts. The two commonest patterns are for the Cairns-Atherton and southern tracts to be occupied by distinctive isolates and for the Cape York and Cairns-Atherton tracts to be populated by forms that have reached a lesser degree of divergence. Examples are : — Species with a distinctive isolate in each of the three main tracts: Opopsitta diophthalma (see Fig. lOh Megaloprepia mag- nifica, Tregellasia capito-leucops. Species with distinctive isolates in the Cairns-Atherton and southern tracts only : Ptilonorhynchus violaceus, Ailuroedus cras- sirostris, Sericornis citreogularis, and 8. magnirostris. Gerygone 406 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY (igata) richmondi has, in addition, an isolate in the Bowen- Mackay tract. Species with isolates in the Cape York and Cairns tracts only : Machaerirhynckus flaviventer, Meliphaga analoga and M. gracilis. In the flycatcher Arses the two tracts are populated hy separate species. The warbler Gerygone palpebrosa has, in addition, a third isolate in the Mackay-Roekhampton section. The barriers breaking up rain forest tracts in Australia, and hence the distribution of rain forest species, have been listed by Tate (1952). They are areas of dry savannah, as follows: — Coen- Cooktown gap (southern Cape York), width 150 miles; Burdekin gap (Townsville-Bowen), 125 miles; Mackay-Roekhampton gap, possibly less complete, about 150 miles; Rockhampton-Miriam- vale gap (near Bundaberg), 100 miles; Toowoomba-Brishane gap, 75 miles. A number of the bird species appear to be absent from the whole section between about Mount Spec, near Townsville. and the Bunya Mountains. In the case of these, the effective dis- tributional barrier is some 400 miles wide. This is almost three times the distance between the Cairns-Atherton and Cape York tracts and explains the greater differentiation of many southern forms. It is obvious that the distribution of rain forest bird species in Australia today can only be explained in terms of greater con- tinuity of this habitat in the past. In particular, the "pockets" between the Cairns area and the Bunya Mountains must for- merly have been larger and more extensive. (b) Monsoon Forest Monsoon forests or pseudo-rain forests form narrow strips aloim- many of the coastal rivers in the drver north and north- west of the continent, an area of summer monsoon rainfall Superficially, they resemble rain forests but differ in their re- stricted extent and relative dryness. Monsoon forests, and the denser mangrove areas that are fre- quently adjacent to them, enable a number of eastern rain forest species to occur in the Northern Territory. There are about 12 of these and, with the exception of only two or three (e.g. Chibia bractcata, Gerygone magnirostris), all are represented by dis- tinctive northwestern isolates. Such species are: Bhipidura rufi frons, Piezorhynchus alecto, Drymodes superciliosa and Poccilo- dryas superciliosa, Lala S > tc - a; © o Iz; <3 rt " tl h£ 1/ IB 5 CD 1=1 c5 «3 io~ - ~s. f. I. X CO I- o o - — ' — 72 . o o - s bo >> r o8£ 3 3 ■*> 3 - — I I— I a — - x ci Q —i Eh i— t H o ■_ ; y. -S a, s M "3 £ -g = .S © "o / — a 3 « "S 03 o 2 ° t- -v X O h " - ■> = * 3 9 s d ^ 8 je tc 2 "O o: ", •- 2 "j= KEAST : BIRD SPECIATION IN AUSTRALIA 463 X *? 3. 3 •£ ii fa fa fa fa .a ft fa 2 o o _i 'S. c 'm 3 ** = t3 m C - H £ «J & fa £ 3 32 re 2 *y *— — » +84 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY — -r. ^ "if* /. - a a soil i5 O 5 "2 b a _ m fci rih«o 0> — a; r = o - 1 - :- - Eh - ■ 73 - - 3 W « .© o 2 ■ ° .5 - — 2 ■ x x M X X X X X O 03 rt o rt — .2 KEAST : BIRD SPECIATION IN AUSTRALIA 465 1* 03 a rt a CD - — 03 & ~ X 'J ^-c 22 ^ p* - a- g OD en rt DC © ^> c 01 O c o ft O ft o 'u a js n r j Ph 33 o B '3 03 rt tli X rt o rt 5 X 03 Q3 ^^ « ~ « t* ft ^ — - j_, p p_, &, j_ c_ a. - p. rt a o 2 /-V Cw in „ t— 1 2 (-c: ,5 tx^S c3 ce 3 "3, *3 cu P- a- 466 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY /. - 43 C m5 fc' X t] '- S3 OS - tad at XXX x ro x ■* cc x x XX XX s - — a: - c Eh X XX X 3 7". 3 ^ u =3 .2 P eS rr >» es a 5 a BO e9 -Z ft ft © 05 as S c a 3 s - — O c5 o ge ft oj ft o v eS >- *- ft c S3 ft oj * .2 ft 06 2 c "S « g< M .2 ,2 ** "«3 3 a k _, rt 00 o ft & 2 ft IS .« X *^ O ft o .2 C8 31 43 - Ph — — fu ft _ — _ ; O = 3 ft E 43 cj *" O as G £ 3 ft ft '^» 99 ft 03 4= ft 03 3 a> o °, 92 OS 3 S £ B * 03 03 OP c* 43 SX fce § o ,2 -3 o •2 £ J< +3 ~ -u 43 «j ' S & © 2 T3 c ft *- O CL,' KEAST : BIRD SPECIATION IKT AUSTRALIA 407 CO i-H r-f ft * ^: ci CJ MM H M M K M M M M M M en tC CJ j-H CO CJ M MM KM M M M K M M M K 05 — i - 3 3 O S Q 2 0) 0J o — 55 3 O to o (3 o s .- & £ .3 to "3 Ph Ph 3 u - x OB s «s 05 3 i- o 00 3 7. 3 W 3 o 3 as 3 3 3 3 05 C> 'J .'2 < di Ph P- Pl. Cm Ph Pi 3 3 Ph 3 .5 ..H -a 3 3 m 3 03 .rt 05 3 ^ 3 3 3 u 3 2: O 3 £ 3 O H -S 05 3 « 3 M .5 O »H 3 « i ^ 3 3 08 2 S 3 45 fc"3 v p, a 3 ° S 35 *"■" c9 2. a> 3 ft s 05 (J3 05 £ tt 05 o o 05 05 S*> >, '— (- c3 3 3 ,3 ft > ,3 Pi W O CU* Ph' Ph" Ph £ fc £ fc fc •J(i,s BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY n3 V a o Q CO w ►J a j S=Jg 0=3 > < - — < M O •_ i < — •— Q H C - M H CO CJ K M x x < p — w H PS o a cd x x x I 3! O 09 GO _ CO mi 5 C •£ (73 3 ft T3 O 3 • is "^ S £ 55 J 3 a> o3 2-5 2 y 3 3 o >. - o E9 £ o 03 3 -tH 3 « i£ CS 3 o .£ E a - .3 "S t ft a3 =5 -c >. -r ^ ^ o. « d a td w a e5 50 ft o 0> > o c 3 3 0> .2 3 >- 3 S r-J t) 3 a =: 00 O £ KEAST: BIRD SPECIATION IN AUSTRALIA 409 .M CJ * * k Jr 1 * x k x x x x Ph h « O E : i ojcj f— c:i A m h ci 5) h S Ph Ph S 09 - — - < M O X, X K - M P a Ph ocjc^grtsg^w ^ 2 S s "2 = = >< ^ 5.5' s s » ! « ^ Hi ^ - •? .2 « 2 2 g = = o 8 £ j? S\S I ff-£ 1 S S 3 «^ *AJ» . § 9 -g £ g f § °!3§qS?.5S"io ft 3 S g f. | |. S A- & 2 "S A | 1 J a § £ S .3 g g Ph 6 D O 'J D J 1-3 S PS S PS 02 Ph S S S S - 0.' Z — £h~ Si 2 . c o =° C '-0 t-» - . S o 5 a a t-. S * 5 a ~ - up CO « e3 S3 a> o C3 U 'oD d H 3 o 3 c 35 _S3 'Z "o be «- S Cm O * 3 .S o o o .s Ch a. C3 cS o _2 § -S 3 * ~ * _ I •* ., r— « :*> T3 B< ai ai cu pC k - * bi o o e 3 ■ 03 "A Cu, cj © <- .3 3 *> a> 3 3 3 o bt > 3 23 05 £ o: 3 £ 3 '►>» £ £0)03 2 « 2. S> g o S3 w O 3 .2 S3 — ft ft EH .1! C 3 Pi S3 S3 « ed s bj ^ K» O 03 3 3 ft * * 3 C8 g 3 ~ - 3 3 o ad Ph Pm Ch Cm CUi KEAST : BIRD SPECIATION IN AUSTRALIA 471 < i— i >—. < i r. I X OJ i rt rl OJ t-i CQ M o OS O I— I u M K tc 05 C es or o r n (J Co S c « h « s a ^ 3 k a & M > o w 05 05 £> 3 c« 15 ft 05 O CS - i- , &5 £ s s ^» a O J= ^ S N •' | 03 ft o3 3 m • a 3 2 5 C o 3 >> *J "H 1 oo ■" C3 .9.2 3 fl o 3 05 « .3 ssssss a ssss6 O C2 ft Qj 3 .s 1 00 15 •< 3 5P 05 O OO oo "a § ° 3 -3 oo oi 22 « 3 S S . .g 05 05 J ei , w - 13 rs ft r! o o i.» 09 05 es o of 476 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY >H Oi SI Zc a o +j • •SB a N X X X XXX H X M N x x y, x x X X bt X >- P3 — ci x oa o: S3 « 3 3 bn to o o d d n 5 - © o ~ ~ © = © rt bx rt rt 4) rt rt 2 7t © .u .3 © - 33 ■s « ■ « © a © P © c 33 ^> •3 5 05 33 be •<- o) cr. — o o 05 r~ rt © 93 S 9 .2 is a P. ^ o T3 C rt 3 f-. rt 3 3 id - « ssssssssss © >- rt c rt S ~ © (h © ^ 03 5 § J5 © rs s .§ rt © rt H rt 03 rt o © ^ it, c ~ cc E © 05 © t. ~ ^ 'C © « t. O .M -C .rt ! © © e •— »— *^- ^3 *-^ w-^ w^ ^^ «•* r KEAST : BIRD SPECIATION IN AUSTRALIA 477 N © o 3 ft -a "5 ^ o U « ^ . o 3 5 © © 3 'E o *- fa re- re - a © ±: 3 ~ _ rr, , . ^ HSs3*lt re ft © ^, 3 ft ..©(-© Cli £« P-, H J re o E --. ft > on 3 re 00 '3d 2 re .5 to S as 3 X - © > re J J J o © ft COB "C — c re © "S re re 6 re 3 - V s X '5 "ft KEAST: BIRD SPECIATION IN AUSTRALIA 479 CO CO to CO 10 01 10 M X X X i-l c3 PH 01 -. rt M Cj 03 S £ 3 q o3 3 bo £ c3 o3 3 3 03 bJ) ._, £ c3 C 3 CD c8 C5 05 £ 0) S q 03 x x co O M 3 o 2 S 5 o >, 3 cr-^ q fi -ri ^ ^ o o o O ii cb 03 u 3 &> "o 3 03 *►> o c3 3 03 rP ft 3 33 o X ° „ O C 03 o> o3 H 3 ri O CO 03 03 3 P CD Z 03 IB CD i O ^3 3 £ 3 !Soo5 2 « is 3 03 "3 CO P-i 03 o o s3 3 o3 3 Pi O O 3 CD CO 480 RI'LLETIN: MUSEUM OF COMPARATIVE ZOOLOGY o J/2 PC < E-h O Eh Eh** ^ o 2 Ph kh r* o i — i Eh CZ2 c 'J ? B-2Z. U H = OS — oc -H rH 01 CM 00 1—1 OD T— I t— -1 rH 00 HH CI CI CO to i— I O t> :i co » -* - 1 i- HfflOiClNIOHMHHC-lM - '8 c CO fa, C J lO CI 00 CI -H -* CI cis o N CO •* N H 00 r-l •-' eg C> CO o ^ 43 B3 T3 3 PL. -tl CM 4) CS O • rH fl - tO CM rH lO CM rH CM CI CM 2* ^ ■-i © 43 03 CD C! rH rH HH O ■* 39 in S «5 Eh O < cS Ph 43 - O O 13 OS — u - — • rH ft o 03 -rH cS s JS 03 J, g -H - o O < § to 43 CS 13 X « KEAST : BIRD SPECIATION IN AUSTRALIA 481 fflg 35 H Eh O TO rV H O 0. j a to 3 a; ~^ .J ■*2sS * a s CO SJ HH rl t- tl CO N I- Pi *i Za 1= to CO * -. : i ci ~ ' CO 2 « — ' 03 T3 00 w .3 M 3 "i i o g -S 3 03 cS .-. M oj T 2u O ^ 3 WOO fH O .J-s r b'^ £ <» •- ? . 3 -^ 2 3"£ 3 oj «h >h 03 T3 ft* «5 T3 .S Oj 3 .3 CS -f- <—* p <; OB 3 << CO 5 >3 2 .2 « W <0 t) 3 i— ( 00 ^ ' CS ^ rrt ^ r-l r-) ., Cv oo «S - CS ^ OS .^^ ri "ti "> : ^ cJ OS,C - o S w a> ft C1 »o m 3 o ^-t 3 ft ■— 5 rH ^ rt s .-2 CS T5 3 O 60 3 'Sd 3 o r— I 03 ^ 2T3 CS O O T3 0) ft.r-. ft cs £< oo oC ° ft CO r3 3 "C - CS -rH CS rH O — O .2 2 +^ *^ • h 03 — oo a> +H 4- ~ 3 43 - &H ■* 3 rS •" y,*- 03 rH ^ 3 3~ ^ r- "H -^t-H H OS 3 ~ 2 t- cS 03 03 •- ,C «tH 3 T3 rH S o y. - 0J u cS rO .s 00 OJ 3.2 cc *> ft OJ 00 f* CJ H ft 3 oo 3 O 5 00 482 BT'LLETIN : MUSEUM OF COMPARATIVE ZOOLOGY i~ H _ o — in <* ^ ^ J h r 1 ^ o o I— I — I i— i I— 1 2K me x 3 O xl = bod x- » O «■> ° — ,_, 4.' •-- r = ~ o< 2 "9 rj - j ; I« z •- ' so- p as « -. ^ - Z M a Ej> = tH 2 Z = O'" i.' .- * foist z'. » to Sri s a 00 00 o 00 io 2 *■ N *" " Ui Tjl « i 1 a x h co ■* O O CI 00 5) « M CO X 00 O Ifi 1— ( i-H OS -# r- Vi i-h r~ o as i- -i O OS i—l 00 © n a :i oc -# i-H •* 00 CI © i.o l> co ci co ci O O CO M» H M H ^ O CI O CO C] -* CO a ® h X f-l n x N CO rH 31 00 CO ■* m © ID o cS ■ 4S r. c3 o o t- 0> a a a a cs cj CO CO CI I- Cl C] C] 00 o C] CI CO CI OS 3a o 3". <§ O Ct CS m h 5 « 02 05 02. § S Q * — KEAST : BIRD SPECIATION IN AUSTRALIA is:; c ~ 'X' u — r. x - H as 3Q 1 9 fa "^ 2 '■ . a c * j • on o> t- a, - w mil s U OS *_' O] I.O — I co ci ■* «i *i; H l> » "ii N r-l CO CI 01 W 01 '.- :i w m -n -* co >* ci «© IB (O W H CI CI 00 00 00 CO in O 00 r-t CO CI H i-l t- # t-; -t 1 CO o -V ■— I i— I CO CO Si -f CO '.O CXI 00 00 -f i ~ «o > - t~ © ci co — < 0CJ CXI co' io «o ci i.o to i-i ©J ao o r-5 co U0 -^ LO o ci IfS 00 r-j rH CO ■* IO CO CXI CO ■— 1 t- 03 «0 -■*' -* 01 01 CI c 5 <° cj -e ■^ "^ r* — ** — X 1* .— o o o o o o o o o o o o o 484 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY 25 O Pn /; Eh o < r. O - W Eh ■/. K < C E .d s led t- Sb £2° 4> •P. at .« n ft 4 ' (H - C *3 02 o_; a <-> • o O 2 o te >>- >• — ' 4) 4-1 - CO 3". c- -f o CO 31 *x> X - I to z. ro IO ITS o o - ) OS ci CO a a -3 to n CD OS 3S 3". O ft T3 S o CO CO 3", to 31 S 1 7 1 CI o up 00 -r - - 1 71 — CI 3V 3". 3. CI o 2 ~ 3 ~ fl c -fl T3 o a> m a, « 02 u a. 02 C3 05 - 31 - A3 "_ - d S • fl A- 02 u fl, ? - tfi o: « O ~ &C . IB fl © <-> v: a +> .fl — X o cS " 3 ££ a? x ^ ci-4 ^tr •^ o ° s^ ^ KEAST: BIRD SPECIATION IN AUSTRALIA 485 Table 8 ISOLATION AND SPECIATION ON OFFSHORE ISLANDS Island Distance offshore (miles) Morphologically Distinctive Isolates Tasmania (southeastern) King Island (southeastern) Kangaroo Island (southern) Recherche Archipelago (southwestern) Abrollhos Islands (western) Dirk Ilartog Island (western) Barrow Island (western) Bernier Island (western) Melville Island (northwestern) 130. Longest water gap. 50 50 from Tasni. 50 from Vict. 5-10 5-20 40 35 20 20 Capricorn Islands (eastern) 25 About 12 that have reached or are approaching the degree of differentiation typical of spe- cies, plus 20 lesser isolates. Minor forms of Melithreptus gularis, M. affinis, Acanthoryn- chus tenuirostris, Acanthiza ewingi, Malurus cyaneus (also on Flinders Island). Total: about 5. Minor forms of Melithreptus atricapillus, Anthochaera chry- soptera, Phylidonyris pyrrhop- tera, Meliornis novaehollandiae, Aeanthorhynchus tenuirostris, Meliphaga leucotis, Estrilda temporalis, Acanthiza pusilla, Stipiturus malachurus, Zonae- ginthus bellus. Total : about 10. Minor form of Sericornis ma- culata. Minor forms of Turuix varia, Sericornis maculata. Malurus leucopterus (also on Barrow Is.). Minor forms of Calamanthus fuliginosus, Sti- piturus malachurus. Eremiornis carter! (minor form). An insular race of Anthus au- stralis has been described from here. Malurus lamberti minor form. Minor forms of Turnix casta nota (?), Coracina papuensis, Dacelo leachii, Meliphaga fla- vescens, Entomyzon cyanotis, Philemon novaeguineae, Myzo mela obscura (?). Total: about 7. Minor form of Zosterops lat- eralis. 4Sf> ni'LLETIX : MUSEUM OF COMI'ARATm; ZOOLOGY Table 9 AVIFAUNAL RELATIONSHIPS BETWEEN AUSTRALIA AND THE MORE REMOTE SURROUNDING ISLANDS Species that have colonized Australia Australian species that have colonized island New Zealand nil Lord Howe Island New Caledonia Timor nil nil 20-22 (Mayr, 1944b) 8-10 in recent times (Falla, 1953), plus many older ele- ments (Meliphagidae, Musciea- pinae). Up to about 10 (see Hindwood, 1940). About 18 (see Mayr, 1945a). About 17 (Mayr, 1944b). New Guinea 66 plus 92 plus. keast: bird speciat10n in australia 487 Table 10 CONTINENTAL AND ARCHIPELAGO SPECIATION. COMPARISON OP NUMBER OP MORPHOLOGICALLY DIFFERENTIATED ISOLATES IN BIRD SPECIES COMMON TO BOTH AREAS Number of Morphologically Number of differentiated Isolates Isolates in Species in southwest Pacific (Fig! 28) Australia Podiceps novae-bollandiae 2 Aviceda subcristata 4 1 Ealiastur sphenurus — — Haliastur indus — — A_ceipiter novae-hollandiae 9 — Acc-ipiter fasciatus — — ( 'iicus approximate — Faleo percgrinus 1 — I'andion haliaetus 1 — Notophoyx novaehollaiidiae — — Demigretta sacra 1 — Nyeticorax caledonicus 2 Butorides striatus 4 5 Dupetor flavicollis 1 — Hotaurus stellaris — — Dendrocygna arcuata 1 — Anas superciliosa 1 — A gibberifrons 1 — Aythya australis — Rallus philippensis 12 1 Rallina tricolor 2 — Porzana tabuensis 1 Poliolimnas cinereus 4 — Amaurornis olivaceus 1 — Porphyrio porphyrin 5 2 Turnix maculosa 1 — 488 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Table 10 (Continued) Species Number of Morphologically Number of differentiated Isolates Isolates in in southwest Pacific (Fig. 28) Australia Tumrs varia Chalcophaps indica Trichoglossus haematodus Eurystomus orientalis Halcyon sanctus Halcyon chloris Coracina lineata C. papuensis C. tenuirostris Lalage leucomela Rhipidura rufifrons R. leucophrys R. fuliginosa R. rufiventris Piezorhynchus alecto Petroica multicolor Pachycephala pectoralis P. rufiventris Artamus leucorhynchus Nectarinaria jugularis Myzomela dibapha Lichmera indistincta Philemon novaeguineae Zosterops lateralis Erythrura trichroa Chibia bracteata Aplonis metallica 1 2 •J 26 6 5 10 5 IS 1 2 5 1 10 34 1 4 1 1 1 6 4 2 1 3 1 1 2-3 1 2-3 1 4-5 2 1 5-7 1 6 TOTAL 201 38-43 KEAST : BIRD SPECIATION IN AUSTRALIA 489 XVI. REFERENCES Amadon, D. 1941. Notes on some Australian birds of prey. Emu, 40:365-384. 1942. Notes on some non-passerine genera. Am. Mus. Novit. no 117;") : 1-11. 1943. Birds collected during the Whitney South Sea expedition. 52. Notes on some non-passerine genera. Am. Mus. Novit., no. 1237: 1-22 1950a. Australian mud nest builders. Emu, 50:123-127. 1950b. The Hawaiian honeycreepers (Aves, Drepaniidae). Bull Am. Mus. Nat. Hist., 95:157-262. 1951. Taxonomic notes on the Australian butcherbirds (family Crac ticidae). Am. Mus. Novit., no. 1504:1-33. Amadon, D. and G. Woolfenden 1952. Notes on the Mathews collection of Australian birds. The order Ciconiiformes. Am. Mus. Novit., no. 1564:1-16. Andrewartha, II. G. and L. C. Bikch 1954. The distribution and abundance of animals. Univ. of Chicago Press, 782 pp. liAKNARD, II. G. 1926. Birds of the Cardwell district, Emu. 26:1-13. Barnard, C. A. and H. G. Barnard 1925. A review of the bird life on Coomooboolaroo Station Duaringa district, Queensland. Emu. 24:252-265. Brown, W. L. and E. O. Wilson 1956. Character displacement. Systematic Zoology, 5:49-64. Browne, W. R. 1945. An attempted post-Tertiary chronology for Australia. Proc. Linn. Soc. N. S. Wales, 70:5-25. BlTTLER, B. E. 1956. Parna-an aeolian clay. Aust, J. Sci., 18:145-151. Cain, A. J. 1955. A revision of Triohoglo&sus haematodus and of the Australian platycercine parrots. Ibis, 97:432-479. Chisholm, A. H. 1951. The story of the scrub-birds. Emu, 51:89-112. Condon, II. T. 1941. The Australian broad-tailed parrots (subfamily Platycercinae). Rec. South. Aust. Mus., 7:117-144. 1951a. Variation in the brown hawk. Emu, 50:152-174. 1951b. Notes on the birds of South Australia: occurrence, distribution and taxonomy. S. Aust. Ornith., 20:26-28. 1954. Remarks on the evolution of Australian birds. S. Aust. Ornith. 21:17-27. Condon, H. T. and D. Amadon 1954. Taxonomic notes on Australian hawks. Rec. South Aust. Mus., 11:189-246. 490 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY < Irocker, R. L. 1959a. Past climatic fluctuations and their influence upon Australian vegetation. Chapter in Biogeography and Ecology in Australia. Pp. 283-290. W. Junk, The Hague, pp. 1-640. 1959b. The plant factor in soil formation. Aust. .lour. Sei., 21 (6a): 180-192. Crocker, R. L. and J. G. Wood 1947. Some historical influences on the development of the South Au- stralian vegetation communities and their bearing on concepts and classification in ecology. Trans. Roy. Soc. S. Aust. ,71:91-130. Darlington, P. J. 1957. Zoogeography: the geographical distribution of animals. John Wiley & Sons, New York, 675 pp. Darwin, C. 1890. The origin of species. 6th ed., John Murray, London, 458 pp. David, T. W. E. 1950. The geology of the Commonwealth of Australia. Ed. W. Browne, London. Vol. 1, 747 pp. Fairbkidge, R. 1953. Australian stratigraphy. 2nd ed. Univ. Western Australia, Perth, 516 pp. Fa i, la, R, A. 1953. The Australian element in the avifauna of New Zealand. Emu. 53:36-46. Fleming, C. A. 1950. New Zealand flycatchers of the genus Petroica Swainson. Paris I and II. Trans. Roy. Soc. New Zealand, 78:14-47; 127-160. Galbraith, I. C. J. 1950. Variation, relationships and evolution in the Pachycephala pec- toralis superspecies (Aves, Museieapidae). Bull. Brit. Mus. (Nat. Hist,, Zool.), 4:133-222. Gardner, C. A. 1944. The vegetation of Western Australia. .lour. Roy. Soc. W. Aust.. 28:xi-lxxxvii. 1959. The vegetation of Western Australia. Chapter in Biogeography and Ecology in Australia. Pp. 275-282. W. Junk. The Hague, pp. 1-640. Gentilli, J. undated. Colored map of Australian climates, Thornwaite Formula, pub lished by the author, University of Western Australia. 1949. Foundations of Australian bird geography. Emu, 49:85-129. (I ill, E. D. 1955. Radiocarbon dates for Australian archeological and geological samples. Aust. Jour. Sci., 18:49 52. I rOLDSCHMIDT, R. 194(1. The material basis of evolution. New Haven. Yale University Press, 436 pp. KEAST: BIRD SPECIATION IN AUSTRALIA 4 ( .)^ Hills, E. S. 1040. Physical geography of Australia. Chapter in The Australian Environment. Pp. L3-22. Com. Sri. Ind. Res. Org., Melbourne. HlNDK, R. A. 10.10. Behaviour and speciation in birds and lower vertebrates. Biol. Revs., 34:85-128. HlNDWOOD, K. A. 1040. The birds of Lord Howe Island. Emu, 40:1-86. HlNDWOOD, K. A. AND E. Mayr 1040. A revision of the striped-crowned pardalotes. Emu, 48:40-07. Howard, H. 1040. A review of the Pleistocene birds of Fossil Lake, Oregon. Car- negie Inst. Washington. Pub., no. 551 : 141-1 Of). Kkakt, A. 1956a. Variation in the Australian Oriolidae. Proc. Hoy. Zool. Soc. N. 8. Wales, 1954-55:19-24. 1956b. Variation in the genus Megalurus. Proc. Roy. Zool. Hoc. N. S. Wales, 1954-55:25-28. 1057a. Variation in the Australian whitefaces (Aphelocephala) Proc. Roy. Soc. N. S. Wales, 1955-56:38-42. 1057b. Variation in the bristle-birds (Dasyornis). Proc. Roy. Zool. Soc. N. S. Wales, 1955-56:43-40. 1957c. Variation in the Australian emu-wrens (Stipiturus.) . Proc. Hoy. Zool. Soc. N. S. Wales, 105.1-50:47-53. 1957d. Variation in the Australian kingfishers (Aves: Alcedinidae). Aust. Mus. Rec, 24:01-72. 1057e. Variation and speciation in the Australian tree creepers (Cli- rnacteris). Aust. Jour. Zool. 5:474-495. 1058a. Variation and speciation in the Australian flycatchers. Rec. Aust. Mus., 24:73-108. L958b. Speciation in the genus Amytornis Stejneger (Passeres: Musci capidae, Malurinae) in Australia. Aust. Jour. Zool., 6:33-52. 1058c. The relationship between seasonal movements and the develop ment of geographic variation in the Australian chats (Epthia- nura Gould and Ashbyia North). Aust. Jour. Zool., 6:53-08. 1058d. The influence of ecology on variation in the mistletoebird (Dicacum liirundinaceum). Emu, 58:195-206. L958e. Seasonal movements and geographic variations in the Australian wood-swallow3 (Artamidae). Emu, 58:207-218. I958f. Infraspecific variation in the Australian finches. Emu, 53:210 246. 1058g. The genus Psophodes Vigors and Borsfield, and ils significance in demonstrating a possible pathway tor I he origin of Eyrean species from Bassian ones. Emu, 58:247 2.15. 10.18h. Geographic variation in the weebill, Smicrornis brevirostris a sedentary species with a continuous range. Aust. .lour. Zool., 6:152-161. 492 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY 1958i. Variation and speciation in the Australian Campephagidae. Aust. Jour. Zool., 6:248-267. 1959a. The Australian environment. Chapter in Biogeography and Ecol- ogy in Australia. Pp. 15-34. W. Junk, The Hague, pp. 1-640. 1959b. Australian birds: their zoogeography and adaptations to an arid continent. Chapter in Biogeography and Ecology in Aus- ralia. Pp. 89-114, W. Junk. The Hague, pp. 1-640. 1959c. The reptiles of Australia. Chapter in Biogeography and Ecology in Australia. Pp. 115-135. W. Junk. The Hague, pp. 1-640. 1959d. Relict animals and plants of the Macdonnell Ranges. Aust. Mus. Mag., 13:81-86. 1959e. The role of islands in evolution. Aust. Mus. Mag., 13:128-132. Keble, B. A. 1947. Notes on Australian Quaternary climates and migration. Mem. Nat. Mus. Victoria, 15:28-81. Lack, D. 1944. Ecological aspects of species-formation in passerine birds. Ibis, 86:260-86. 1947. Darwin's finches. Cambridge University Press, 208 pp. 1954. The natural regulation of animal numbers. Oxford Univ. Press, 343 pp. Leeper, G. W. 1949. Climates of Australia. Chapter in The Australian Environment. Pp. 23-24. Com. Sci. Ind. Res. Org., Melbourne. LUNDELIUS, E. 1957. Additions to knowledge of the ranges of Western Australian mammals. West. Aust. Nat., 5:173-182. Lynes, II . 1930. Review of the genus Cisticola. Ibis, (12 ) 6:673 pp., 20 pis. Mack, G. 1934a. Notes on the genus Orthonyx. Emu, 34:1-7. 1934b. A revision of the genus Malurus. Mem. Nat. Mus. Melbourne, 8:100-125. 1936. A systematic revision of the Australian thornbills. Mem. Nat. Mus. Victoria, 10:86-120. Main, A. R., A. K. Lee and M. J. Littlejohn 1958. Evolution in three genera of Australian frogs. Evolution, 12: 224-233. Mathews, G. M. 1910-1927. The birds of Australia. 12 vols. II. P. and G. Witherby, London. Mayr, E. 1934. Notes on the genus Petroica. Am. Mus. Novit., no. 714:1-19. 1937. Notes on the genus Sericwnis, Gould. Am. Mus. Novit., no. 904: 1-25. 1940a. Notes on New Guinea birds. Am. Mus. Novit. no. 1056:1-12. 1940b. Notes on Australian birds. The genus Lalage. Emu, 40:111-117. KEAST : BIRD SPECIATION IN AUSTRALIA 493 1941. List of New Guinea birds. Am. Mus. Nat. Hist, N.Y., 260 pp. 1942. Systematic* and the origin of species from the viewpoint of n zoologist. Columbia Univ. Press, New York, 334 pp. 1943. Notes on Australian birds II. Emu, 43:3-17. 1944a, The birds of Timor and Sumba. Bull. Am. Mus. Nat. Hist., 83:123-194. 1944b. Timor and the colonization of Australia by birds. Emu, 44:113 130. 1945a. Birds of the Southwest-Pacific. MacMillan Co., New York, 316 pp. 1945b. Notes on the birds of northern Melanesia. Part I. Am. Mus. Novit., no. 1294:1-12. 1945c. Wallace's Line in the light of recent zoogeographic studies. Science and Scientists in the Netherlands, Indies, Surinam and and Curacao. New York, pp. 241-250. Also: Quart. Rev. Biol., 19:1-14, (1944). 1947. Ecologic factors in speciation. Evolution, 1:263-288. 1948. Geographic variation in the reed warbler. Emu, 47:205-210. 1949. Notes on the birds of northern Melanesia, Part 2. Am. Mus. Novit,, no. 1417:1-38. 1950a. Speciation in birds. Proc. 10th Inter. Ornith. Congress, Stock- holm, Pp. 91-131. 1950b. Taxonomic notes on the genus Neositta. Emu, 49:282-291. 1951. Notes on some pigeons and parrots from Western Australia. Emu, 51:137-145. 1953a. Geographic and individual variation in the shrike-tit, Falcun- cuius frontatus. Emu, 53:249-251. 1953b. Taxonomic notes on Oreoica gutturalis. Emu, 53:252-253. 1954a. Notes on Australian whistlers. Am. Mus. Novit., no. 1653:1-22. 1954b. Changes of genetic environment and evolution. Chapter in Evo- lution as a process. Ed. J. Huxley. Allen and Unwin, London, Pp. 157-180. Mayr, E. and D. Amadon 1951. A classification of recent birds. Am. Mus. Novit,, no. 1496:1-42. Mayr, E. and S. Camras 1938. Birds of the Crane Pacific expedition. Zool. Ser., Field Mus. Nat. Hist., 20:453-473. Mayr, E. and K. Jennings 1952. Geographic variation and plumages in the Australian bower birds (Ptilonorhynchidae). Am. Mus. Novit., no. 1602:1-17. Mayr, E. and M. Moynihan 1946. Evolution in the Rhipidura rufifrons group. Am. Mus. Novit., no. 1321:1-21. Mayr, E. and D. L. Serventy 1938. A review of the genus Acanthiza, Vigors and Horsfiehl. Emu, 38:245-292. 1944. The number of Australian bird species. Emu, 44:33-40. 494 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY Matr, E. and R. Wolk If*'?). The south-west Australian races of the spotted scrubwren, Seri- cornis maculatus. West Aust. Naturalist., 4:66-70. McGill, A. R. 1948. A distributional review of the genus Neositta. Emu, 48:33-52. McMlCHAEL, D. AND T. IREDALE 1959. The land and freshwater Mollusca of Australia. Chapter in Biogeography and Ecology in Australia. Pp. 224-245. W. Junk. The Hague. Mees, G. F. 1957. A systematic review of the lndo-Australian Zosteropidae (part I). Zool. Verhandel. No. 35, 1-204. Meise, W. 1931. Zur Systematic der Gattung Gerygone. Novit. Zool., 36:317-379. Miller, A. H. 1951. An analysis of the distribution of the birds of California. Univ. Calif. Publ. Zool., 50:531-644. 1955. Concepts and problems of avian systematica in relation to evo- lutionary processes. Recent Studies in Avian Biology. Ed. A. Wolf son. University of Illinois Press, Urbana. Pp. 1-22. Moreau, R. E. 1948. Ecological isolation in a rich tropical avifauna. Jour. Anim. Ecology, 17:113-126. North, A. J. 1901-1914. Nests and eggs of birds found breeding in Australia and Tasmania. 4 vols. Australian Museum, Sydney. Paramanoy, S. J. 1959. Zoogeographical aspects of the Australian dipterofauna. Chapter in Biogeography and Ecology in Australia. Pp. 164-191. W. Junk. The Hague. Peters, J. L. 1934. Check list of birds of the world. Cambridge, Mass. vol. 11, 401 pp. Pre scott, J. A. 1931. The soils of Australia in relation to vegetation and climate. 2nd ed. Council Sci. and Ind. Res. Bull., 52:1-82. 1944. A soil map of Australia. Com. Sci. Ind Res. Org. Bull. no. 177. Rand, A. L. 1936. Meliphaga analoga and its allies Am. Mus. Novit., no. 872:1-23. Ripley, S. D. 1942. A review of the species, Anas castanea, Auk, 59:90-99. Scrivenor, J. B., T. H. Burkill, M. A. Smith, A. S. T. Corbet, H. K. Airy shaw, P. W. Richards and F. E. Zeuner 1943. A discussion of the biographic division of the lndo-Australian Archipelago, with criticism of the Wallace and Weber Lines and of any other dividing lines and with an attempt to obtain uni- formity in the names used for the discussions. Proc. Linn. So?. London, 154th Sess., 120-165. KEAST : BIRD SPECIATION IN AUSTRALIA 495 sf.rvf.xty, I). L. 1951. The evolution of the chestnut-shouldered wrens (Malurus). Emu, 51:113-120. 1953. Some speeiation problems in Australian birds. Emu, 53:131-14."). Sf.rventy, D. L. and H. M. Whittell 1951. Birds of Western Australia. 2nd ed. Paterson Brokensha, Perth, 345 pp. Sharland, M. 1945. Tasmanian birds. Angus & Robertson, Sydney, 175 pp. Spencer, B. 1898. Report on the work of the Horn scientific expedition to central Australia. Narrative. Dulau & Co., London, 220 pp. Stresemann, E. 1927-1934. Handbueh der Zoologie. Sauropsida. Aves. Berlin. Tate, G. H. II. 1952. Mammals of Cape York peninsula, with notes on the occurrence of rain forest in Queensland. Results of the Archbold expedi- tions. No. 66. Bull. Am. Mus. Nat. Hist., 98:567-616. Taylor, J. K. 1949. Soils of Australia. Chapter in The Australian Environment, p. 35. Comm. Sci. Ind. Org., Melbourne. Tindale, N. B. 1949. New Satyridae of the genus Oreixenica from South Australia and New South Wales, together with notes on the recent climate of Southern Australia. Rec. S. Aust. Mus., 9:143-155. Wakefield, N. A. 1958. The yellow-tufted honeyeater with a description of a new sub species. Emu, 58:163-193. Wood, J. G. 1949. Vegetation of Australia. Chapter in The Australian Environ- ment. Pp. 77-96. Com. Sci. Ind. Res. Org., Melbourne. 1959 The phytogeography of Australia (in relation to radiation of 'Eucalyptus, Acacia, etc.). Chapter in Biogeography and Ecology in Australia. Pp. 291-302. W. Junk, The Hague, pp. 1-640.