Cladistic analysis of 157 characters of definitive plumages and soft parts, natal plumages, tracheae, and nontracheal skeletons of 59 Anatini (sensu Livezey 1986) provided a phylogenetic hypothesis of high consistency (CI = 0.71, excluding unique autapomorphies) and resolution (tree completely resolved except for two nested trichotomies, a trichotomy within the northern-hemisphere mallards, and the tentative placements of two poorly known species of Anas). Major phylogenetic inferences (lists of three or more taxa are in order of increasing relatedness) include the following: (1) monophyly of the tribe is weakly demonstrated; (2) the tribe comprises three subtribes--Cairineae (Cairina, Pteronetta, Aix), Nettapodeae (Chenonetta, Nettapus), and Anateae (all other genera); (3) subtribe Anateae comprises three "supergenera" (each comprising two genera)--Amazonetta (A. brasiliensis and Callonetta leucophrys), Lophonetta (L. specularioides and Speculanas specularis), and Anas (genera Mareca and Anas); (4) the genus Mareca or wigeons includes six species--capensis, strepera, falcata, sibilatrix, penelope, and americana; and (5) the large genus Anas comprises two weakly supported subgroups or cohorts, the first of which includes two subgenera (mallards [Anas] and blue-winged ducks [Spatula]) and the second includes four subgenera (Australasian teal [Nesonetta], pintails [Dafila], Holarctic teal [Querquedula], and spotted teal [Punanetta]). Other findings include that a sister-relationship exists between Anas sparsa and other mallards, that subgroups of the "true" mallards are largely congruent with biogeographic subdivisions (northern-hemispheric, African, and South Pacific), that the four species of shoveler are monophyletic, that a sister-relationship between the speckled teals (flavirostris and andium) and brown pintails (georgica, acuta, and eatoni) exists, and that A. querquedula is the sister-group to the green-winged teals (formosa, crecca, and carolinensis) and not closely related to the blue-winged ducks. Supplementary data and related theory indicated that (1) interspecific interfertility is a poor indicator of relationship, (2) the phylogenetic species concept provided the most practical definition of terminal taxa, (3) a majority of groups within the Anatini originated in the southern hemisphere, (4) body size, sexual size dimorphism, and egg size are strongly constrained phylogenetically, whereas sexual dichromatism and clutch size are less predictable, (5) biparental attendance of broods is primitive within Anatini, and (6) characters of definitive plumages, natal plumages, tracheae, and skeletons had similar consistencies but attained maximal utility at different levels within the phylogeny, which indicates different rates of character evolution within the tribe. Received 19 July 1990, accepted 4 December 1990.

Museum of Natural History, Dyche Hall, University of Kansas, Lawrence, Kansas, 66045-2454 USA TRIBE Anatini (dabbling ducks; sensu Livezey ! 986) comprises all typical surface-feeding ducks and the smaller "perching ducks," which were formerly grouped under the tribal name "Cairi- nini" (Delacour and Mayr 1945) or included in the subfamily Plectropterinae (Phillips 1922). These are perhaps the most familiar waterfowl and together include roughly one third of the extant species of Anseriformes. Bock and Far- rand (1980) listed Anas as one of the largest avian genera, yet they conservatively recog- nized only 36 species. Current taxonomic groupings within the Anatini, like other sub- groups of waterfowl, derive principally from the classic work by Delacour and Mayr (1945) and, to a lesser extent, earlier investigations (Salvadori 1895; Phillips 1922, 1923, 1925, 1926; Peters 1931; Kuroda 1939; and Boetticher 1942). The taxonomy and inferred evolutionary "af- finities" of dabbling and perching ducks pro- posed by Delacour and Mayr (1945), based on comparisons of behavioral and external simi- larities, were expanded by Delacour (1954, 1956, ! 959, 1964). Subsequent research on the system- atics of dabbling ducks used comparisons of be- havior and morphology (e.g. Boetticher 1952; Lorenz 1941, 1951-1953; Johnsgard 1960a, b, 1961a, b, 1962, 1965; Kaltenhauser 1971; Scherer and Hilsberg 1982) and largely confirmed the phenetic groupings defined by Delacour and Mayr (1945). Important changes proposed at two major taxonomic levels included the relation- ships of several problematic species to the tribe, and relationships among the species within the Anatini. Most of the taxonomic revisions recommend- ed subsequent to Delacour and Mayr's (1945) proposal pertained to generic or tribal mem- bership. Johnsgard (1960c) determined that the Ringed Teal ("Anas" leucophrys) differed suffi- ciently from typical Anas to be returned to the monotypic genus Callonetta. In a comprehen- sive study of waterfowl osteology, Woolfenden (1961) recommended that the "perching ducks" (Tribe Cairinini, sensu Delacour and Mayr 1945, but exclusive of Plectropterus) be merged into the "dabbling ducks" (Tribe Anatini). Johns- gard proposed to re-assign the Pink-headed Duck (Rhodonessa caryophyllacea; 1961a) and Marbled Teal (Marmaronetta [Anas] angustiros- tris; 1961c) to the pochards (Tribe Aythyini). Johnsgard (1961a) also advocated the inclusion of the Crested Duck (Lophonetta specularioides) within the Anatini rather than among the shel- ducks (Tribe Tadornini) as proposed by Dela- cour and Mayr (1945). Salvadori's Duck (Sal- vadorina [Anas] waigiuensis) was included in Anas by Delacour and Mayr (1945), but earlier au- thors placed it in a separate subfamily Merga- nettinae with Merganetta (Salvadori 1895, Phil- lips 1926), and others retained it as a monotypic genus of problematic relationships (Mayr 1931, Peters 1931, Kear 1975). Several other species included as aberrant members of the Anatini by Delacour and Mayr (1945) subsequently have had varied taxonomic treatments. These include the Blue Duck (Hymenolaimus malacorhynchos), Pink-eared Duck (Malacorhynchus membrana- ceus), and Freckled Duck (Stictonetta naevosa). Hymenolaimus was studied in detail (Kear 1972), and Stictonetta was later removed from the Anat- inae on the basis of morphological and behav- ioral comparisons (see Johnsgard 1960a, 1961a, 1962). The unique Torrent Duck (Merganetta ar- mata) was placed in the monospecific tribe Mer- ganettini (Delacour and Mayr 1945) but consid- ered to be closely related to Anas (Niethammer 1952). Merganetta was included by Delacour (1956) and Johnsgard (1961a, 1965) within the Anatini, although Johnsgard (1978) later re- turned it to its own tribe. Eldridge (1979) con- cluded that Merganetta is most similar ethol- ogically to Tadorninae. These taxonomic revisions were adopted in large part by Johnsgard (1979) in the latest edi- tion of the "Checklist of Birds of the World," although tribes within the Anatinae were omit- ted. In a recent phylogenetic analysis of Recent genera of Anseriformes, Livezey (1986) con- cluded that (1) Stictonetta and Plectropterus are representatives of monotypic, pre-anatine branches of subfamilial rank, (2) Sarkidiornis, Malacorhynchus, Hymenolaimus, and Merganetta are unusual members of the Tadorninae, (3) Marmaronetta and Rhodonessa are members of the Aythyini, and (4) the relationships among the remaining members of Cairinini and Ana- tini could not be determined from osteology and therefore were merged into the single, pos- sibly paraphyletic tribe Anatini. Livezey (1986) omitted Salvadorina from this analysis, but MI- kovsk (1989) reexamined the partial skeletons of S. waigiuensis described by Mayr (1931) and concluded that the species may be more closely related to Hymenolaimus and Malacorhynchus than to Anas. Relationships among dabbling ducks, partic- ularly within the genus Anas, have received comparatively little study. Special attention was given relationships within Anas (Delacour and Mayr 1945; Lorenz 1951-1953; Johnsgard 1961a, 1965). These studies were based largely on com- parisons of behavior (and to a lesser extent, plumage) and none attempted to compile com- parative data for even a majority of species. All (except perhaps Lorenz 1951-1953) were phe- netic in nature (i.e. groupings of species were based on assessments of overall similarity ir- respective of possible polarities of characters). A pronounced "tradition" of taxonomic se- quences of dabbling ducks developed (bench- mark works include Salvadori 1895; Phillips 1922-1926; Peters 1931; Delacour and Mayr 1945; Johnsgard 1961a, 1979), which reflected a con- sensus of roughly "primitive-to-advanced" in- tratribal relationships. These classifications pri- marily reflect changing taxonomic conventions and secondarily indicate differences in percep- tions of relationships among Anatini. Varia- tions in species names aside, important trends include a reduction in numbers of species of Anatini (sensu Livezey 1986, excluding Salva- dorina) recognized (65 by Salvadori 1895, 64 by Phillips 1922-1926, 60 by Peters 1931, 46 by Delacour and Mayr 1945 and by Johnsgard 1961a); a reduction in number of genera rec- ognized (17 by Salvadori 1895, 10 by Phillips 1922-1926, 13 by Peters 1931, 7 by Delacour and Mayr 1945, and 8 by Johnsgard 1961a); and an increase in the numbers of presumed "natural" groups within Anas, a change that largely re- flected the "Groups 1-14" of Delacour and Mayr (1945) and the clusters depicted by Johnsgard (1961a) or considered "superspecies" by Johns- gard (1979). Molecular studies of Anseriformes have con- tributed significant, largely confirmational in- formation on probable relationships (Scherer and Sontag 1986). A number of biochemical studies have included more than one species of Anatini. Among these are compositional com- parisons of uropygial secretions (Jacob and Gla- set 1975, Jacob 1982); electrophoretic studies of proteins from a diversity of tissues (Sibley and Ahlquist 1972, Brush 1976, Numachi et al. 1983, Oates et al. 1983, Patton and Avise 1985); im- munological comparisons of blood plasma (Bottjer 1983); restriction endonuclease analysis of mitochondrial DNA (Kessler and Avise 1984, Avise et al. 1990); and DNA hybridization (Madsen et al. 1988). None of these studies, however, included samples from >14 species of Anas or > 19 species of Anatini (sensu Livezey 1986), few considered problematic species, and none employed cladistic methodologies. Branching diagrams in these studies usually were based on simple UPGMA clustering of taxa based on pairwise distances. For the resul- tant branching diagrams to represent phylo- genetic relationships, such clustering algo- rithms require the additional assumption of equal rates of molecular evolution in all lin- eages (Wiley 1981, Hillis 1987). Moreover, some of the, studies were hampered by (I) small num- bers of loci compared (e.g. I0 loci in Numachi et al. [1983] and 26 in Patton and Avise [1985]), (2) data matrices that lack substantial propor- tions of pairwise distances and have undesira- ble metric properties (Madsen et al. 1988), (3) a substantial dependence on traditional classifi- cations in investigational design and inferences (Bottjer 1983, Patton and Avise 1985), or (4) a reliance on traditional, phenetic classifications of fossil species for estimates of divergence times and evolutionary rates (Patton and Avise 1985, Madsen et al. 1988). I present a phylogenetic (cladistic) analysis of the Anatini from 157 polarized characters of the skulls, syringeal bullae, natal plumages, de- finitive plumages, and soft parts of 59 taxa. I discuss relative consistencies and evolutionary rates of character groups, ecological and bio- geographical correlates, and relationships of traditionally problematic taxa and closely re- lated tribes, and I propose an annotated Lin- nean classification of the Anatini. MATERIALS AND METHODS Taxa included.--All 88 specific and subspecific taxa of Recent Anatini were included for analysis as mem- bers of the ingroup, although skin specimens of downy young, tracheae, and (especially) skeletons were not available for all taxa. For the derivation of trees, I combined taxa with identical character-states, which reduced the number of analytical units to 59. These analytical combinations merged taxa typically con- sidered to be subspecies. I defined subspecies for prac- tical purposes as lineages that differed only quanti- tatively (often in intensity of plumage colors or size) from conspecific taxa. These differences probably re- flect spatially separated samples of clinal variation. This practice produced the nontraditional combina- tion of Anas fulvigula and A. diazi (while maintaining the qualitatively distinguishable A. platyrhynchos and A. rubripes), and the elevation to species rank of a number of taxa commonly treated as subspecies (A. wyvilliana, A. laysanensis, A. oustaleti, A. zonorhyncha, A. albogularis, A. chlorotis, A. eatoni, A. andium, A. car- olinensis, and A. puna). Several other taxa are either of problematic tribal assignment (e.g. Salvadorina, Mer- ganetta) or are "basal," Anatini-like members of other tribes of Anatinae (e.g. Polysticta, Marmaronetta, Het- eronetta) and were included for ancillary comparisons. These latter genera and several others (Sarkidiornis, Tadorna, Somateria) were included as members of the outgroup for determination of polarities. Specimens examined.--I examined study skins of downy young (mostly age-class Ia, brightly patterned with no plumaceous feathers) and adults, skeletal ma- terial (particularly skulls), and syringeal bullae (most mounted with tracheae) of all species of Anatini ex- cept for the following: Anas oustaleti (downy young not examined, but described by Phillips [1923]), A. chlo- rotis and A. aucklandica (downy young not examined, but described and illustrated in Delacour [1956, 1964]); A. melleri (skeleton and syrinx not available), A. al- bogularis (skeleton and syrinx not available), and A. bernieri (skeleton, syrinx, and downy young not known). I included a series of study skins of adults of all species, and for a majority of species, samples of three or more were studied for characters of the skull, trachea, and natal plumage. The collections of study skins of adults and downy young at the Na- tional Museum of Natural History (NMNH) and American Museum of Natural History (AMNH), and the skeletal collections at the NMNH and the Uni- versity of Kansas Museum of Natural History (KUMNH), were especially important. I received a generous loan of skins of downy young and mounted tracheae from the Wildfowl Trust (WT), Slimbridge, England. Analysis of some characters was facilitated by study of several published references: Phillips (1922-1926), Delacour and Mayr (1945), Delacour (1956, 1964), Rip- ley (1957a, b), Huey (1961), Woolfenden (1961), Johnsgard (1965, 1978), Oring (1968), Palmer (1976), Soothill and Whitehead (1978), Todd (1979), Weller (1980), Hosking and Kear (1985), Madge and Burn (1988). Data on body mass, sexual dimorphism, dis- tributions, and reproductive parameters were com- piled from these references and others listed by Live- zey (1990). Definition and analysis of characters.--A total of 157 characters that had two or more discrete states were included in the analysis (Appendix 1). Based on out- group comparisons, I divided characters into a single primitive state (for the Anatini) and one or more de- rived states. To confirm the monophyly of terminal taxa, representative autapomorphies were compiled (but not included in the analysis; a list is available from the author on request). Key references for char- acter descriptions were as follows: Humphrey and Parkes (1959) for definitive plumages, Humphrey and Clark (1964) and Warner (1971) for the syrinx, and Butendieck and Wissdoff (1982) for the skull. Several skeletal characters given by Woolfenden (1961) for subgroups of the Anatini (especially Amazonetta and Callonetta) were excluded because of intraspecific variation and problematic definition of states. Behav- ioral characters of Anatini have been studied intense- ly (Delacour and Mayr 1945; Lorenz 1951-1953; Mc- Kinney 1953, 1965; Delacour 1956; Johnsgard 1961a, 1962, 1965), but they were excluded because of dif- ficulties in definition of character states and numer- ous missing data. Derivation of phylogenetic trees.--Searches for most- parsimonious topologies (trees requiring the fewest character changes) were made with "global" branch- swapping algorithms with "MULPAR$" (MAXTREE = 250) in the PAUP program (Swofford 1985). Both procedures search for the "shortest" tree(s) possible and permit the storage of a number of equally par- simonious topologies, but neither guarantees the de- termination of the shortest possible tree(s). Exhaus- tive search algorithms (e.g. branch-and-bound) were not employed because of the prohibitively great com- putational expenditure necessitated by data sets of this size. The 16 multistate characters (those with two or more derived states) were analyzed as unordered, with the exception of the meristic, ordinal counts of rectrices (character 27). To find preliminary topolo- gies within reasonably short computing times, I re- placed several unknown syringeal character states in A. bernieri with those shared by other Australian teal (A. gibberifrons, A. castanea, A. chlorotis, and A. aucklan- dica); similarly, several natal character states of A. bernieri were hypothesized to be those shared by A. gibberifrons and A. albogularis. I also described the to- pological differences of subsequent analysis with missing-data codes replaced. All characters were as- signed unit weight. Trees were rooted using an "hy- pothetical ancestor" (vector of plesiomorphous char- acter states), which was prepared from polarities inferred from outgroup comparisons. (See Appendix 3 for an annotated data matrix.) Theory and meth- odological rationales for phylogenetic analysis are given by Wiley (1981) and Swofford (1985). All com- putations were made on the IBM mainframe com- puter at the University of Kansas. Phylogenetic classification.--The hierarchical group- ings within phylogenetic trees can be summarized in Linnean classifications (Wiley 1981). Parageneric taxa--supergenera, subgenera, and infragenera--were erected as needed to conserve phylogenetic infor- mation and based on synonomies presented by Phil- lips (1923), Brodkorb (1964), and Wolters (1976). RESULTS Trees found.--Under the constraint of char- acter states hypothesized (as detailed above) in the Madagascan Teal (A. bernieri), inferred trees had consistency indices of 0.71 and lengths of 250 (inclusion of representative autapomor- phies inflated the consistency index by 0.05). Only three unresolved sections of the phylog- eny were indicated (Fig. 1). The first poorly resolved segment involved the relationships among three subgenus-groups in the genus Anas (sensu stricto). Three topological variants were found: (1) the groups (Nesonetta, Dafila, and Pun- anetta-Querquedula) as a trichotomy; (2) Dafila as the sister-group to a clade that comprises Dafila and Punanetta-Querquedula; and (3) Punanetta- Querquedula as the sister-group to a clade com- posed of Dafila and Nesonetta. The second poorly resolved section involved the relationships among four entities: the poorly known A. ber- nieri; the relatively similar A. gibberifrons and A. albogularis; and a resolved clade that includes A. castanea, A. chlorotis, and A. aucklandica (Fig. 1). Four equally parsimonious topological variants were found for these four groups: (1) A. bernieri, A. albogularis, and A. gibberifrons composing a grade paraphyletic to the "reddish teal" (A. cas- tanea, A. chlorotis, and A. auklandica); (2) A. ber- nieri as the sister-group to a trichotomy involv- ing A. albogularis, A. gibberifrons, and the clade of "reddish teal"; (3) a trichotomy involving A. bernieri, a clade of A. albogularis and A. gibberi- frons, and the clade of "reddish teal"; and (4) A. bernieri as the sister-group to A. gibberifrons + A. albogularis, and together these three as the sister- group to the "reddish teal." Given the poor res- olution of these two segments of the tree, the included relationships are considered indeter- minate and depicted as trichotomies (Fig. I). The third poorly resolved segment involved two nested trichotomies including A. diazi, A. ful- vigula, and the clade composed of A. platyrhyn- chos and its Pacific relatives (Fig. I). All other parts of the phylogeny in the constrained anal- ysis were invariant. Replacement of the hypothesized character states in A. bernieri with missing-data codes re- suited in 81 trees (CI = 0.715), in which A. ber- nieri is associated with the "first cohort," either as a monotypic branch of a trichotomy with or as the sister-group to the other two member clades. This placement was based entirely on the questionably homologous reddish foot-col- or of A. bernieri. If this character is coded as not homologous to the orange feet of mallards and shovelers, then A. bernieri was placed with equal parsimony in a diversity of "basal" positions among the subgenera throughout the genus Anas. Exclusive of A. bernieri and the two tri- chotomies described above, the tree based on the unconstrained analysis (Fig. 1) was invari- ant. Tribal monophyly and "perching ducks."-- Monophyly of the Anatini is only weakly sup- ported by three character changes of which two were lost by reversal in subsequent groups. The "perching ducks" comprises two clades para- phyletic to the remainder of the Anatini, hence- forth referred to as Anateae (Fig. 2). The first and more primitive of these was the clade com- posed of two species of Cairina and its mono- typic sister-genus Pteronetta, and two species of Aix. The two species of Aix have been consid- ered sister-species by most authorities in recent decades (e.g. Delacour and Mayr 1945, Delacour 1959, Johnsgard 1965, Numachi et al. 1983). The second clade of "perching ducks," a group more closely related to the true dabbling ducks (Fig. 2), comprises two genera--Chenonetta (one species) and Nettapus (three species). The close relationship of these two genera, and the para- phyly of the smaller "perching ducks" differs from earlier reviews (e.g. Delacour and Mayr 1945; Delacour 1959; Johnsgard 1961a, 1965, 1978). Particularly divergent previous treat- ments include the placement of Chenonetta in the monotypic subfamily Chenonettinae (Sal- vadori 1895, Phillips 1922) or inclusion of Che- nonetta within the true geese, subfamily Anser- inae (Peters 1931). In contrast, Chenonetta and Aix were depicted as sister-genera by Johnsgard (1961a, 1978). Basal Anateae.--Two monotypic genera of small, uniquely patterned waterfowl of South America--Amazonetta and Callonetta--were found to be sister-genera and together formed the sister-group to the rest of the subtribe An- ateae (Fig. 2). Boetticher (1952) and Johnsgard (1960c, 1961a, 1965, 1978) included both genera among the "Cairinini," but Delacour (1964) re- tained them within the Anatini as formerly as- signed (Delacour and Mayr 1945, Delacour 1959). Delacour (1964) also resisted the hypothesis of close relationship between Amazonetta and Cal- lonetta supported by Derscheid (1938), Verhey- en (1955), Johnsgard (1960b), and Woolfenden (1961). I found another couplet of taxonomically con- troversial, relatively primitive, Neotropical dabbling ducks--Lophonetta specularioides and "Anas" specularis--to be sister-species, and these composed the sister-group to the true dabbling ducks (Fig. 2). With the exceptions of Delacour and Mayr (1945), Delacour (1954, 1964), Boet- ticher (1958), and Woolfenden (1961), who placed Lophonetta within the shelducks (Tador- ninae), systematists have considered both forms to be somewhat aberrant members of the An- atini (e.g. Johnsgard 1961a, 1965; Brush 1976; Livezey 1986). The proposal of close relation- ship between specularioides and specularis has re- ceived less support by previous studies (e.g. De- lacour and Mayr 1945). Although the two species were listed next to one another by Salvadori (1895), Phillips (1923), and Peters (1931), a sis- ter-relationship was depicted by Johnsgard (1961a). Morphological evidence that indicates a close relationship was cited by Johnsgard (1965). Wigeons and allies.--The remaining members of the tribe make up two major clades. Ranked as genera, they are the wigeons and close rel- atives (Mareca, 6 species) and the typical dab- bling ducks (Anas, 40 species). The wigeons are typified by a combination of primitive syringeal bullae and uniquely derived natal and defini- tive plumages. In addition to the typical wi- geons (M. sibilatrix and the sister-species M. americana and M. penelope), the unique Cape Teal (M. capensis) and two transitional forms (M. strepera and M. falcata) were included within this clade (Fig. 2). The Cape Teal (or Cape Wi- geon, locally; Delacour 1956) formerly was in- HYPOTHETICAL Chononotta j'ubata ANCESTOR ____ r- Nottapus auritus [  At coromandel/anus '  N. pulchellu$ Callonotta /oucophrys Lophonotta spoculorJoidos  peculanas specularis  MoroCa caponsis ANAS t A. galer/culata  Cairina mo$chata C, scutulata ' Ptoronotta hartlaub/ Ancestor ? ? ? ? ? m o (c) N, coromande/ianu$ ig. 2. Detailed topology of ph3/41ogen3/4 of Anatini exclusive of genus Ana$; taxonomy follows Appendix 2. Characters are explained in Appendix 1; autapomorphies are given in parentheses following corresponding taxa. cluded with either the "spotted teal" (e.g.A. versicolor; Delacour 1956) or the "green-winged teal" (e.g.A. crecca; Johnsgard 1961a, 1965, 1979). Mareca capensis combines a natal plumage and several features of definitive plumages shared uniquely with typical wigeons with a moder- ately derived syringeal bulla and a diversity of autapomorphies of definitive plumage and soft parts. The Gadwall (M. strepera), long known for a number of wigeon-like characters of behavior, plumage, and biochemistry (Delacour and Mayr 1945; Boetticher 1952; Delacour 1956; Johnsgard 1961a, 1965, 1978; Kessler and Avise 1984), proved to be a mosaic of a mallard-like natal plumage, several wigeon-like characters of de- finitive plumages, and a syrinx and soft parts that combine features of both groups with an assortment of unique autapomorphies. The close relationship of the Falcated Duck (M. falcata) to the typical wigeons was corroborated (four characters of adult plumage, two natal charac- ters). Delacour and Mayr (1945) allied falcata most closely with the Baikal Teal (A. formosa), but most previous workers noted some "affin- ity" between falcata and the wigeons or strepera (e.g. Delacour and Mayr 1945, Delacour 1956, Boetticher 1958, Johnsgard 1978). Johnsgard (1961a) depicted a sister-relationship between falcata and the typical wigeons, a topology at variance with subsequent commentary (Johns- gard 1965). The terminal clade of typical wi- geons--wherein the Chilo Wigeon (M. sibila- trix) is the sister-species to the northern- hemisphere "superspecies" that comprises M. penelope and M. americana (Johnsgard 1979)--is especially well supported here (Fig. 2). A sister- relationship between these two species was in- tuited earlier by Johnsgard (1961a). Typical dabbling ducks (Ana$).--Approximate- ly two thirds of the tribe Anatini are members of the large, comparatively derived clade of Fig. 1. Phylogeny of dabbling ducks (Anatini). Poorly resolved relationships within the lower clade of Anas and among austral teal shown as trichotomies; position of A. bernieri is tentative. Taxonomy follows Appendix 2. "typical" dabbling ducks. The group includes several of the most widespread, numerous, and familiar species of waterfowl in the world. My analysis defines two weakly supported sub- clades within the group, which I informally refer to as "cohorts" (Fig. 3). The first cohort comprises two major subgroups, the mallards (subgenus Anas; 14 species) and the blue-winged ducks (subgenus Spatula; 6 species). The sec- ond cohort comprises three primary clades, which are shown as a trichotomy because of the poor resolution of this segment of the tree. It includes the Australasian Teal (subgenus Neso- netta; 6 species), the pintails (subgenus Dafila; 7 species), and a two-parted clade that com- prises the Holarctic teal (subgenus Querquedula; 4 species) and the spotted teal (subgenus Puna- netta; 3 species). Mallards.--Monophyly of the mallard com- plex is supported by six characters of definitive and natal plumages, colors of soft parts, and syrinx (Fig. 3). The first dichotomy within the complex separated the African Black Duck (Anas sparsa), a species that combines primitive ana- tomical characters with atypical (evidently aut- apomorphic) behavioral attributes, from a spe- close clade of more-typical mallards. Other investigators recognized the "distinctness" of A. sparsa (e.g. Delacour and Mayr 1945, Boet- ticher 1952, McKinney et al. 1978), and several even considered the species to be among the most primitive members of the Anatini (Dela- cour 1956; Johnsgard 1961a, 1965, 1978), al- though Johnsgard (1979) listed A. sparsa with the more typical mallards. The sister-group of A. sparsa includes a sub- clade of the (largely) northern-hemisphere mal- lards and another, southern-hemisphere group that includes South Pacific mallards and two remaining African members (Fig. 3). Relation- ships among the northern mallards are incom- pletely resolved. The American Black Duck (A. rubripes) is the sister-group to a trichotomy that includes the Mottled Duck (A. fulvigula), Mex- ican Duck (A. diazi), and a reasonably well sup- ported clade that includes the Mallard (A. platy- rhynchos) and three closely related insular forms--the Marianas Duck (A. oustaleti) and the Hawaiian sister-species A. wyvilliana and A. lay- sanensis. The northern mallards have been con- sidered closely related for decades, an hypoth- esis supported by biochemical comparisons (Kessler and Avise 1984, Patton and Avise 1985). Several authorities believe that most or all of the continental forms are conspecific and most also include the Pacific forms within an en- larged, polytypic A. platyrhynchos (e.g. Delacour and Mayr 1945; Boetticher 1952; Delacour 1956; Johnsgard 1960e, 1961a, 1961d, 1965, 1967, 1978, 1979). Others retained species rank for some members (e.g. Phillips 1923, Moulton and Wel- let 1984, Hepp et al. 1988). The African mallards, Yellow-billed Duck (A. undulata) and Meller's Duck (A. melleri), consi- tute one of two clades included in the sister- group of the northern-hemisphere mallards (Fig. 3). Delacour and Mayr (1945: 21) stated that each A. undulata and A. melleri "stands alone" within the mallard complex but listed them next to one another within their "Group 9." There is a rel- atively long tradition of listing A. undulata next to A. sparsa (Salvadori 1895, Phillips 1923, Peters 1931, Johnsgard 1979). Johnsgard (1961a, 1965) concluded that A. melleri was very closely re- lated to A. platyrhynchos, but later he (1978) con- cluded that available information did not per- mit a precise assessment of the relationships of A. melleri. Johnsgard (1978, 1979) listed A. un- dulata and A. melleri consecutively. The sister-group of the two African forms comprises the four species of South Pacific mal- lards (Fig. 3). Within the Pacific complex, the Philippine Duck (A. luzonica) is the sister-group to a terminal clade in which the Pacific Gray Duck (A. superciliosa) is the sister-group to the spot-billed ducks, here considered to comprise two species (A. poecilorhyncha and A. zonorhyn- cha). With few exceptions (Boetticher 1952, Johnsgard 1961a), recent investigators inferred the close relationship among the South Pacific mallards. This led most taxonomists to merge zonorhyncha with poecilorhyncha (e.g. Johnsgard 1979), to combine both zonorhyncha and super- ciliosa within poecilorhyncha (Delacour 1956; Johnsgard 1961a, 1965, 1978), or (less frequent- ly) to consider all four South Pacific forms (in- cluding the distinctive luzonica) as subspecies of a single species (Delacour and Mayr 1945; opposed by Ripley 1951). Blue-winged ducks.--The first cohort also in- cludes the clade that comprises the blue-winged teal and shovelers (Fig. 3). The two teal are sis- ter-species, with A. discors significantly more aut- apomorphic (uniquely derived) than A. cyanop- tera. The four species of shoveler constitute the sister group to the blue-winged teal. Within the  A. sparsa (126a-b) - A. platyrhynohos   J g. A. wyv/ll/ana      A. Iysnenss (28b-o,29b-o)   I - A. rubr/pes J   A. poeilorhynh   A. onorhynh      ' ' ' ' A ynopter      -- A. sm/thU , , , , A. pltle (0b-a) , , , , , I. A. ANAS ..... " A. rhynhot/s   A. bern/ei (96-b *)     A. gibber/frons (Ho- ) ? ? ? ?  A. hlorot/s  ?    A. ulnd/ (5o-) A. lbogulr/s     A. erythrorhynh    A. flWrostr/s (88o-b, 93o-b) , ?  % %    A. andium   ..... eatoni __            . arolnen  . formosa   A. querquedula -          A. versiolor -       ?  . puna        A. hottentot -,-,- Fig. 3. Detailed topology of phylogeny of the genus Anas; position of A. bernieri tentative (see text). Characters are explained in Appendix 1; autapomorphies are given in parentheses following corresponding ta. Character changes marked by asterisks are mutually contradictor, and alternative arrangements can support several distinct, equally parsimonious topologies within the trichotomy. shovelers, A. smithii is the sister-group to a clade that comprises A. platalea, A. rhynchotis, and A. clypeata, within which the Australasian (A. rhynchotis) and Northern (A. clypeata) shovelers are sister-species (Fig. 3). The blue-winged teal and shovelers have been considered to be closely related by most 20th- century taxonomists, an idea supported by re- cent biochemical comparisons (Kessler and Avise 1984, Patton and Avise 1985). However, the relationships inferred here differ in several ways. First, the Garganey (A. querquedula), tra- ditionally grouped with the blue-winged teal, is shown (Fig. 3) to be more closely related to the green-winged teal (e.g.A. formosa, A. crecca) and (to a lesser degree) to the spotted teal (e.g. A. versicolor). Second, monophyly of the shov- elers was confirmed, contrary to the suggestions of polyphyly by Delacour and Mayr (1945) and Delacour (1956). Specifically, the Red Shoveler (A. platalea) and Australasian Shoveler (A. rhyn- chotis) are less closely related to the superficially similar Cinnamon Teal (A. cyanoptera) and Blue- winged Teal (A. discors), respectively, than they are to each other and to the remaining shovelers (Fig. 3). The relationships proposed here also imply that the Cape Shoveler (A. smithii) is the "most primitive" member of the shovelers. Most extant taxonomic sequences (Delacour and Mayr 1945; Delacour 1956; Johnsgard 1965, 1978, 1979) and the tree depicted by Johnsgard (196 l a) sug- gest instead that A. platalea "links" the shov- elers with the blue-winged teal. Australasian teal.--Five or six species of small, grayish or reddish-brown teal (if the poorly known A. bernieri is included) compose one of the three clades in the second "cohort" of Anas (Fig. 3). Three species--A. bernieri, A. albogularis, and A. gibberifrons--probably form a poorly re- solved, basal grade paraphyletic to a well-sup- ported clade of three, variably reddish-brown teal. The latter species--Chestnut Teal (A. cas- tanea), Brown Teal (A. chlorotis), and Flightless Teal (A. aucklandica)--are limited in distribu- tion to Australia, New Zealand, and the Auck- land and Campbell islands, and show a pro- nounced trend toward smaller size, reduced sexual dichromatism, and flightlessness (Live- zey 1990). Members of this clade have been considered closely related by most taxonomists since De- lacour and Mayr (1945). This concensus includ- ed a close relationship between the poorly known A. bernieri and other austral teal, es- pecially A. gttberifrons and A. albogularis (e.g. Sal- vadori 1895; Phillips 1923; Peters 1931; Dela- cour and Mayr 1945; Verheyen 1955; Johnsgard 1961a, 1965, 1978, 1979), although most taxon- omists also listed "Anas" capensis with or near this group. Several authorities discussed or rec- ommended the conspecific or superspecies sta- tus of A. bernieri and A. gibberifrons (Boetticher 1952; Delacour 1956; Johnsgard 1965, 1979), and most authorities considered some or all of the Australasian "reddish" teal to be conspecific (e.g. Ripley 1942; Delacour and Mayr 1945; Boetti- cher 1952; Johnsgard 1978, 1979; Dumbell 1986). Pintails.--The second major clade within the second "cohort" of Anas includes the pintails (Fig. 3). Anas bahamensis and A. erythrorhyncha form the first of three subgroups, species con- sidered closely related by virtually all 20th-cen- tury taxonomists (e.g. Salvadori 1895; Phillips 1923; Peters 1931; Delacour and Mayr 1945; Johnsgard 1965, 1978, 1979). The second sub- clade of pintails includes two monophyletic groups (Fig. 3): two teal-sized species (A. fiavi- rostris and A. andium), and three species of brown pintails (A. georgica, A. acuta, and A. eatoni). The inference of close relationship between the "pale-cheeked" and brown pintails has re- ceived widespread support, except for Delacour and Mayr (1945), who suggested that the "pale- cheeked" species are more closely related to the "spotted teal" (e.g.A. versicolor), Cape Teal ("A." capensis, herein moved to Mareca), and Marbled Teal (Marmaronetta angustirostris, now placed in Aythyini; Livezey 1986). Anas fiavirostris (in- cluding A. andium) was allied with A. crecca by Delacour and Mayr (1945), Lorenz (1951-1953), Boetticher (1952), Delacour (1956), and Johns- gard (1961, 1965, 1978, 1979). The Brown Pintail (A. georgica), here found to be the sister-group to A. acuta and A. eatoni, was placed "nearer to" the superficially similar A. fiavirostris by several taxonomists (e.g. Salvadori 1895, Delacour and Mayr 1945), but Johnsgard (1965, 1978, 1979) indicated a close relationship between A. acuta (including A. eatoni) and A. georgica. Anas fia- virostris and A. andium have been considered closely related for decades and, since Delacour and Mayr (1945), most authorities have treated them as a single species. The sister-species A. acuta and A. eatoni have been considered con- specific in recent decades (Delacour and Mayr 1945; Johnsgard 1965, 1978, 1979), but Stahl et al. (1984) supported species-status for both. Holarctic and spotted teal.--Two well-support- ed clades of teal compose the third clade in- cluded within the second "cohort" of Anas (Fig. 3). The larger of these includes the Holarctic teal, a northern-hemispheric group composed of the Garganey (A. querquedula) and its sister- group the green-winged teal (A. formosa and the "superspecies" including A. crecca and A. car- olinensis). The finding that A. querquedula is the sister-group to the green-winged teal, support- ed by three characters of definitive and natal plumages, opposes the traditional A. querquedula association with the blue-winged ducks (Sal- vadori 1895; Phillips 1923; Peters 1931; Dela- cour and Mayr 1945; Boetticher 1952; Verheyen 1955; Delacour 1956; Johnsgard 1961a, 1965, 1978, 1979). The classification of A. querquedula with the blue-winged ducks rested largely on the gray forewing of the species, a character that differs from the less extensive, powder-blue forewing of Spatula and that evidently was de- rived from the light gray wing-coverts of the green-winged and spotted teals. A close rela- tionship between the Baikal Teal (A. formosa) and the green-winged teal "superspecies" has been hypothesized by all 20th-century taxon- omists except Delacour and Mayr (1945), Boet- ticher (1952), and Delacour (1956), who consid- ered formosa to be an "isolated" form of uncertain taxonomic position. Spotted teal, the distinctive southern-hemi- sphere sister-group of the Holarctic teal, in- clude the Hottentot Teal (A. hottentota), and the superspecies that comprises the Silver Teal (A. versicolor) and Puna Teal (A. puna) (Fig. 3). Al- though some writers failed to recognize that A. hottentota was a member of this well-defined group (Salvadori 1895, Phillips 1923, Peters 1931), this relationship was supported by De- lacour and Mayr (1945) and subsequent taxon- omists. The practice of listing A. querquedula be- tween A. versicolor and the blue-winged ducks (Salvadori 1895; Heinroth 1911; Lorenz 1951- 1953; Delacour 1956; Johnsgard 1961a, 1965, 1978, 1979) is not in accord with my results. This tradition apparently reflects the phenetic assessment of synapomorphies and symple- siomorphies between A. querquedula and the "spotted teal," and superficial (convergent) similarities between A. querquedula and the blue- winged ducks. DISCUSSION Poorly resolved relationships.--My results con- stitute a significant clarification of the generic relationships of Anatini presented earlier (Live- zey 1986), but monophyly of the tribe remains only weakly demonstrated. Two of the three supportive characters had consistency indices of 0.50, one (character 1) characterized the basal grade Cairineae + Nettapodeae and the other (character 101) also supported (by reversal) the clade Cairina + Pteronetta. The single character supportive of the tribe (character 128, postor- bital stripe in natal plumage) is not completely unambiguous. Although it is absent in most "allies" placed in other tribes (Sarkidiornis, Mer- ganetta, Malacorhynchus, Hymenolaimus, and Marmaronetta), it is suggested in several others (Plectropterus, Salvadorina, and Heteronetta). In addition, a few other character states of these genera resembled those found within Anatini (particularly Anateae). These include variably metallic and bordered wing specula (Merganet- ta, Salvadorina, Polysticta, and Heteronetta) and a moderately derived condition of cranial fora- mina (Heteronetta). In all of these "allied" gen- era (except perhaps Salvadorina), there are other characters that indicate alliance with other tribes (Livezey 1986). Presumably the few characters shared with the Anatini are convergent. Sal- vadorina appears, on the basis of its (tadornine) carpal wing-spurs and peculiar, evidently de- rived wing specula, to be most closely related to Merganetta, which in turn is related to Hy- menolaimus by osteological synapomorphies (Livezey 1986) and a dark, vertical supraorbital stripe in natal plumage (shared also with Het- eronetta; pers. obs., Kear 1972). A close relation- ship among Salvadorina, Merganetta, and Hy- menolaimus was inferred earlier by Kear (1975), but this hypothesis must remain tentative until skeletal comparisons are possible. The two basal members of the genus Mare- ca--M. capensis and M. strepera--presented "conflicts" between definitive, natal, and syrin- geal characters, and require further investiga- tion (Fig. 2). Also, the two "cohorts" of Anas are too poorly documented for formal recognition, but the determination of relationships among the included (and well supported) subgenera is important for biogeographic and ancillary evo- lutionary inferences (see below). Relationships among the basal austral teal (A. bernieri, A. albogularis, and A. gibberifrons) are in- adequately understood because of incomplete data for A. bernieri and apparent character con- flicts between A. albogularis and A. gibberifrons. The most urgent anatomical needs for further study of Anatini concern three species endemic to the Indian Ocean: downy young and tracheae for A. bernieri, tracheae for A. albogularis, and tracheal specimens for A. melleri. Relationships among North American mallards have been studied intensively (Phillips 1912; Delacour 1956; Johnsgard 1960d, 1961d) but remain in- adequately resolved. Of particular concern are the delimitation of species within the complex (Morgan et al. 1976, Ankney et al. 1986, Hepp et al. 1988, Avise et al. 1990) and the conser- vation of genetically "pure" populations of A. rubripes in the face of increased hybridization with A. platyrhynchos (Johnsgard 1967, Heus- mann 1974, Johnsgard and Di Silvestro 1976, Brodsky and Weatherhead 1984, Ankney et al. 1987). Anas oustaleti, probably extinct, shows ex- traordinary plumage variability, wherein some specimens resemble the South Pacific A. super- ciliosa and others show characteristics of the Holarctic A. platyrhynchos. Most ornithologists consider A. oustaleti to be a subspecies of A. platyrhynchos or a hybrid swarm (Delacour and Mayr 1945, Yamashina 1948, Delacour 1956, Amadon 1966, Johnsgard 1978). The hybrid- swarm hypothesis is validated by the observa- tion that characteristics of hybrid waterfowl frequently are not simple "intermediates" or mosaics of those of parental species (Scherer and Hilsberg 1982, Gillespie 1985). Detailed cla- distic analysis of A. oustaleti should permit a test of its possibly hybrid origin and reticulate phy- logeny (Humphries 1983, Wagner 1983). Monophyly of terminal taxa was demonstrat- ed unambiguously for all except A. diazi, A. ful- vigula, A. oustaletL A. gibberifrons, and A. eatoni (list available on request). Although the nom- inate subspecies of A. gibberifrons is character- ized by a unique dorsal inflation of the frontal bones (Ripley 1942), the character evidently does not unite the entire species. Reduction of sexual dichromatism distinguishes A. eatoni from its continental sister-species A. acuta. Reduced di- chromatism is evident in insular populations of many anatids (Weller 1980), including several Anatini (A. wyvilliana, A. laysanensis, A. aucklan- dica,) and the extinct, possibly specifically dis- tinct M. (strepera) couesi (Ripley 1957a, b). Hybridization, species concepts, and phylogenetic inference.--Frequency of interspecific hybrid- ization (in the wild or in captivity) and the fertility of hybrid offspring have been weighed heavily by traditional systematists as criteria for the delimitation of species and for the assess- ment of relationships among species and higher taxa of waterfowl (e.g. Delacour and Mayr 1945; Lorenz 1951-1953; Delacour 1956, 1959; Sibley 1957; Johnsgard 1960b, 1961a, 1961d, 1965, 1968a, 1968b, 1978). In spite of an almost uni- versal capacity for hybridization among Anser- iformes (Scherer and Hilsberg 1982), such in- formation has led to a diversity of inferences. Hybridization between captive A. fiavirostris and wild A. crecca has "proved" a close interspecific relationship (Delacour 1964: 341). Infertility of hybrids between "Anas" leucophrys and Ama- zonetta brasiliensis was used in "excluding" a close relationship between the two (Delacour 1964: 343). The unique chromosomal number of Aix galericulata (Yamashina 1952) was interpreted by Delacour (1959: 99) as the explanation for the "strange" absence of interspecific hybrids in Aix, and a demonstration that karyotypic characters are not reliable taxonomic criteria. Phylogenetic inferences based on hybridiza- tion and sterility are (at least implicitly) based on the concept of "isolating mechanisms," an idea central to the widely accepted "biological species" concept (sensu Mayr 1969), in which species are conceived of as groups of inter- breeding, natural, reproductively isolated pop- ulations. The importance of "isolating mecha- nisms" in the formation and maintenance of species of waterfowl has been widely accepted (e.g. Sibley 1957; Johnsgard 1960d, 1963; but see McKinney 1970, 1975; McKinney et al. 1990), but it is supported only by circumstantial evi- dence and often contradictory reasoning. For example, Johnsgard (1963: 531) concluded that "This remarkable capacity for hybridization [among anatids] indicates that isolating mech- anisms must be operating effectively if species are to retain their integrity under natural con- ditions. Since genetic isolation is practically ab- sent in the Anatidae, other isolating mecha- nisms must, of course, have evolved to take their place. Of these, the most significant appear to be behavioral differences and various morpho- logical (plumage and soft part) specializa- tions .... "With respect to Anatini, however, Johnsgard (1963: 537) stated, "Thus, isolating mechanisms are least well developed in the group, in spite of the fact that courtship displays and male plumages tend to be elaborate and diversified." Johnsgard (1963: 539) reasoned that the frequency of hybrids among species of Anas "... appears to be the combined result of the great amount of natural sympatry [cited earlier (p. 538) as a distributional circumstance con- ducive to the evolution of isolating mecha- nisms], the relatively close relationships of all the species of Anas, and the very uniform pre- copulatory behavior of all the species in this genus [i.e., the failure of precopulatory isolat- ing mechanisms to evolve]." Clearly, if isolating mechanisms have been selected in waterfowl, then this selection must have been of low in- tensity, or the genetic-behavioral capacity of anatids to evolve and respond to these mech- anisms is singularly limited. The latter seems particularly improbable in light of the diversity of breeding plumages and capacity for recog- nition of mates characteristic of anatids (Butch- er and Rohwer 1989). An alternative model is to consider "isolating mechanisms" to be incidental effects, not the causes, of interspecific phenotypic divergence (West-Eberhard 1983, Paterson 1985) that re- suits from sexual selection or divergent natural selection (McKinney 1970, Selander 1972). As such, it predictably would fail to preclude in- terspecific mating under a variety of circum- stances. "Intraspecifically selected" (instead of isolation-oriented) phenotypic characters func- tion in selection of mates and facilitate initia- tion of breeding (Selander 1972, West-Eberhard 1983, Andersson 1986). These characters would be subject to intense selection in anatids in which tertiary sex-ratios typically favor males (Bell- rose et al. 1961, Breitwisch 1989), and they would be influenced by other life-history factors such as relative parental investment by males, ter- ritoriality, and predation (Trivers 1972, Endler 1978, Baker and Parker 1979, Butcher and Roh- wer 1989). A recently proposed alternative to the "bio- logical species" concept is the "recognition con- cept" of species, in which species are defined (intrinsically) as populations of organisms that share a common, functionally adaptive system of fertilization (Paterson 1985, 1988; see also West-Eberhard 1983). The "recognition con- cept" was anticipated by students of waterfowl, at least with respect to proximate mechanisms, by an investigational emphasis placed on mech- anisms of "species recognition" (Sibley 1957; Dilger and Johnsgard 1959; Hailman 1959; Klopfer 1959; Johnsgard 1960d, 1961d, 1963, 1968b; Schutz 1965; Williams 1983). Both con- cepts of species hinge--in fundamentally dif- ferent ways--on the interfertility of popula- tions. The first concept focuses on "isolation" or historico-adaptive partitioning of a common system of fertilization. The second emphasizes the internal mechanisms interpreted as adap- tations for its maintenance through refinement of fertilization systems. Species concepts based on inferred or presumed patterns of gene flow and models of speciation are important for the development of evolutionary theory, but often they offer no practical (specimen-based) means for species delimitation. For purposes of phylogenetic inference, in- terfertility is uninformative because it repre- sents the retention of primitive "recognition systems" and capacity for interbreeding (i.e. it is a symplesiomorphous condition). A third al- ternative, the phylogenetic species concept, de- fines species as the smallest diagnosable (i.e. definable by qualitative, apomorphic characters) monophyletic group of organisms (reviewed by Cracraft 1983, 1988; McKitrick and Zink 1988) and permits the comparatively simple defini- tion of terminal lineages in analyses (regardless of consideration of allopatry or fertilization sys- tems) based on the observable effects of evolu- tionary divergence. Recognition of phyloge- netic species permits the investigation of evolutionary problems that tend to be superfi- cially considered with the merging of evolu- tionary entities into (phenetically defined) "bi- ological" species (e.g. relationships within North American and Hawaiian mallards or among bas- al Australasian teal). Consequently, I adopted the phylogenetic species concept for data cod- ing and the cladistic analyses, and for the clas- sification proposed below. A phylogenetic classification of Anatini.--Based on the phylogenetic relationships inferred herein (Fig. 1), I propose a Linnean classifica- tion of Recent dabbling ducks (Appendix 2). Recognition of subtribal and subgeneric taxa permitted the incorporation of additional phy- logenetic information in the classification not possible within the traditional tribe-genus sys- tem. The limitations of simple binomial classi- fication for Anas led Delacour (1956: 19) to con- clude that "... Either 14 genera, or only one, must be recognized." Although subspecies are not considered use- ful in the phylogenetic species concept (Mc- Kitrick and Zink 1988), recognized taxa are list- ed parenthetically (Appendix 2) following those species considered "polytypic" to clarify the content of species (e.g. populations included within A. versicolor and A. puna) and to indicate traditionally distinguished populations which may, with further study, be usefully elevated to species status. The second point is particu- larly important for widespread species charac- terized by comparatively great variation (e.g.A. cyanoptera), for poorly known, possibly diag- nosable insular isolates (M. [strepera] couesi, A. [aucklandica] nesiotis), and for terminal taxa not shown to be monophyletic (e.g.A. gibberifrons). Biogeographical patterns.--The distributional ranges of Anseriformes have been well-known for decades. Conversely, biogeographic infer- ences have been mostly scattered anecdotal speculations concerning "centers of origin," "relictual" distributions, and possible radia- tions of groups or "ancestral stock," and have been made in the absence of an explicit hy- pothesis of phylogeny (e.g. Ripley 1957b; Wel- ler 1964; Johnsgard 1978; Murton and Kear 1975, 1978; Kear and Murton 1976). A tabular overlay of distributional information for my proposed classification permitted more explicit infer- ences concerning the historical biogeography of Anatini (Table 1). Four of six supergenera and eight of eleven genera are limited to the southern hemisphere. Within the genus Anas, only two subgenera (Nesonetta and Punanetta) are strictly limited to the southern hemisphere, although infragenera within two other subgen- era are so distributed (Table 1). Like the genus Mareca, several subgroups of Anas include members from both hemispheres (subgenus Anas, infragenera Pterocyanea, Spatula, Paecilo- nitta, and Dafila). Only four groups of subge- neric or higher rank are limited to the northern hemisphere. These are the genus Aix of Cairi- neae, the grade that comprises the subgenera Chaulelasmus and Eunetta of Mareca, and the sub- genus Querquedula of Anas (Table 1). The pre- ponderance of southern-hemispheric clades, particularly among the early branches (Netta- podeae, and supergenera Cairina, Amazonetta, and Lophonetta), strongly suggest that the An- atini originated in the southern hemisphere, as inferred for the order Anseriformes (Livezey 1986). Southern origins are indicated throughout the tribe (Fig. 1, Table 1). The tree originates with the Asian-African Cairineae (except for the more northern Aix), followed by the Australasian and African Nettapodeae and the strictly South American grade that comprises the supergenera Amazonetta and Lophonetta. Within the super- genus Anas, the genus Mareca forms an African branch (M. capensis) and a largely Holarctic sis- ter-clade of wigeons and allies. The genus Anas has five major branches, of which one is Aus- tralasian (subgenus Nesonetta), three are char- acterized by African or South American, or both, basal members (subgenera Anas, Spatula, and Dafila), and one comprises groups of Afro-Neo- tropical (subgenus Punanetta) and Holarctic dis- tributions (subgenus Querquedula). The typical mallards (subgenus Anas exclusive of A. sparsa) show pronounced geographical partitioning, with northern-hemisphere, African, and South Pacific clades (Fig. 3, Table 1). Departures from this general pattern are the Neotropical distri- bution of M. sztilatrix within the otherwise Hol- arctic clade Chaulelasmus + Eunetta + Mareca; the Australasian distribution of A. rhynchotis, the sister-species to the North American 4. cly- peata; and the insular isolation of A. laysanensis and A. wyvilliana (infragenus Anas) in the equa- torial Pacific Ocean, and A. eatoni (infragenus Dafila) in the Indian Ocean. At least the last three were probably founded by wayward mi- grant flocks (Delacour 1956, Weller 1980). Morphological correlates.--Several physical characteristics not used in inferring relation- ships within the Anatini showed strong phy- logenetic patterns (Table 1). Body size, defined by three discrete intervals of mean body masses, varied little within genera of Cairineae and Nettapodeae, and (to a lesser degree) within Mareca. In the genus Anas, body-size diversity occurs within most subgenera. A majority of the size variation in the subgenus Anas is due to the dwarfism of the Pacific isolates (A. wyvil- liana, A. laysanensis). In several other subgenera, the ranges in body mass reflect phylogeneti- cally constrained variation among constituent clades (e.g. between Pterocyanea and Spatula, and between Dafilonettion and Dafila). Sexual size di- morphism, as measured by three intervals for ratios of mean body masses of males and fe- males, was associated positively with mean body masses of species (r = 0.54, n = 52). That is, size dimorphism tended to be greater in large spe- cies than in small species (Table 1). Exceptions include the disproportionately small dimor- phism of Chenonetta and some members of the infragenera Polionettea and Paecilonitta, and the disproportionately great dimorphism of A. aucklandica (subgenus Nesonetta). Sexual dichromatism, either sexual differ- ences in color or pattern, or both, in adults in definitive alternate (breeding) plumage, is less easily quantified. To permit broad comparisons with the tribe, I established three categories of sexual dichromatism: class I, qualitative (i.e. sex- T^ILE 1. Biogeographical, morphological, and reproductive correlates of phylogeny of Anatini. Classification follows that proposed in Appendix 2. Codings are explained in footnotes. Sexual Sexual Biogeographic Body size di- dichro- Egg Clutch Taxonomic group area a mass  morphism c matism a mass e size f Subtribe Cairineae AS, AF, NA L-M L-M II M-L M-L Supergenus Cairina AF, AS L-M L-M I-If M-L M-L Genus Cairina AF, AS L L II L L Genus Pteronetta AF M M II M M Supergenus Aix AS, NA M L I M L Genus Aix AS, NA M L I M L Subtribe Nettapodeae AU, AF S-M S I-If S-M M-L Supergenus Chenonetta AU S-M S I-If S-M M-L Genus Chenonetta AU M S I M L Genus Nettapus AU, AF S S II S M-L Subgenus Nettapus AF S S II S M Subgenus Cheniscus AU S S II S L Subtribe Anateae global S-L S-L I-Ill S-L S-L Supergenus Amazonetta NE S S I-If S M Genus Amazonetta NE S S II S M Genus Callonetta NE S S I S M Supergenus Lophonetta NE M M II M-L S Genus Lophonetta NE M M II M S Genus Speculanas NE M M II L S Supergenus Anas global S-L S-L I-Ill S-L S-L Genus Mareca HA, NE, AF S-M S-L I-Ill S-M M Subgenus Notonetta AF S S Ill S M Subgenus Chaulelasmus HA M M I M M Subgenus Eunetta PA M L I M M Subgenus Mareca HA, NE M M I-If M M Genus Anas global S-L S-L I-Ill S-L S-L Subgenus Anas HA, AF, SP, S-L S-L I-Ill S-L S-M HI Infragenus Melananas AF L L Ill L S Infragenus Anasg HA (HI) M-L M-L I-Ill M-L S-M (S-M) (S-M) (11) (S-M) (S-M) Infragenus Polionetta SP M-L S-M III M M Infragenus Afranas h AF M M III M M Subgenus Spatula global S-M S-M 1-11 S-M M-L Infragenus Pterocyanea NA, NE S S I S L Infragenus Spatula global M S-M 1-11 M M-L Subgenus Nesonetta i AU, AS S-M S-L I-II S-L S-L "Infragenus Virago" AU, AS S S II S M Infragenus Nesonettai AU M (S) M (L) I (11) M-L (S) S-L (S) Subgenus Dafila NE, HA, AF S-M S-M 1-111 S-M S-M Infragenus Paecilonitta AF, NE M S II M M Infragenus Dafilonettion NE S S II S S Infragenus Dafila HA, NE M S-M I, Ill M M Subgenus Querquedula HA S S I S M Infragenus Querquedula PA S S I S M Infragenus Nettion HA S S I S M Subgenus Punanetta NE, AF S-M M II S M Infragenus Punanetta NE S-M M II S M Infragenus Micronetta AF S -- II S M  Region: AF = Africa, AS = Asia, AU = Australia, HA = Holarctic, HI = Hawaiian, NA = Nearctic, NE = Neotropic, PA = Palearctic, SP  South Pacific. b Mean body mass: S = small (< 500 g), M = medium (500-1,000 g), L = large (> 1,000 g); data taken primarily from Madge and Burn (1988). c Mass dimorphism ratio, (males)/(females): S = small (1.00-1.10), M = medium (1.11-1.20), L = large (> 1.21). a Chromatic differences: I = qualitative, II = quantitative, III= absent.  Mean egg mass (Rohwer 1988): S = small (25-40 g), M = medium (41-60 g), L = large (>61 g). Mean clutch size (Rohwer 1988): S  smalI (3-6), M = medium (7-9), L  large (> 10).  Parenthetical codes refer to "Horizonefta" (Anas wyvilliana and A. laysanensis). h Reproductive data for A. melleri not available.  Poorly known A. bernieri not included.  Parenthetical codes refer to A. aucklandica. es differ significantly in pattern as well as de- gree); class II, quantitative (i.e. sexes differ pri- marily in intensity or conspicuousness, but un- derlying pattern is essentially the same); and class III, obsolete or entirely absent. Despite problematic assignments of several species with "intermediate" levels (e.g. Nettapus, Lophonetta, M. strepera, A. wyvilliana, A. fiavirostris), several patterns emerged (Table 1). First, 22 of 38 spe- cies (excluding A. oustaleti) are characterized by class-I dichromatism and 27 show class-II di- chromatism. Together, 49 (84%) species of An- atini are sexual dichromatic. Second, sexually monochromatic species (class III) are limited to the supergenus Anas and include members of Mareca and the subgenus Anas. Third, class-I dichromatism characterizes all member species in only three named groups: the genus Aix, sub- genus Querquedula, and infragenus Pterocyanea. Sexual dichromatism within the subgenus Anas is exceptionally variable and ranges from the monochromatic condition of southern- hemisphere species and an insular isolate (A. laysanensis) to the conspicuous dichromatism of A. platyrhynchos. "Reduced" sexual dichroma- tism characterizes insular Anatini (especially Anas) and species of the southern hemisphere (Table 1). This pattern has been interpreted tra- ditionally as the adaptive loss of "isolating mechanisms" in insular communities that lack phenotypically "confusable" congeners (Ripley 1957b, Sibley 1957, Johnsgard 1963, Lack 1970, Weller 1980). Alternatively, West-Eberhard (1983) suggested that the comparatively dull plumages of insular waterfowl reflect a reduced intensity of sexual selection related to the pro- tracted pair-bonds and greater parental invest- ment by males (Weller 1980). This hypothesis is supported by the strong association between biparental care of young and sexual monochro- matism in Anatidae (Kear 1970). Another pos- sibility is that the dull, juvenal-like plumages of insular anatids are nonadaptive characters developmentally linked to a selectively fa- vored, paedomorphic reduction of the pectoral limb (Livezey 1989a, 1990). Behavioral patterns.--Although homologies and polarities of behavioral characters were in- adequately established for inclusion in the phy- logenetic analysis, evolutionary trends in re- productive behavior of Anatini can be documented. Age of sexual maturity ranges from 2 yr in Cairina (characteristic of Tadorninae, Plectropterus, Sarkidiornis and evidently primi- tive) to ! yr in all other Anatini (Johnsgard 1961a, 1965; Kear 1970). Cairina is also unique among Anatini in its promiscuous (polybrach- ygynous) mating system (Johnsgard 1961a) while monogamy is typical of dabbling ducks (McKinney 1985). Biparental attendance of broods characterizes most members of the out- group (Kear 1970), Pteronetta (Kear 1970), Net- tapodeae (Frith 1967), Amazonetta (Phillips 1923), Callonetta (Brewer !989), Lophonetta (Buitron and Nuechterlein !989), Speculanas (Kear !970), some Mareca (Weller 1968, Siegfried 1974), and to varying degrees the members of most subgen- era of Anas (Weller 1968, Siegfried 1974, Mc- Kinney !985, Norman and McKinney 1987, Mc- Kinney and Brewer 1989). Cairina and (to a lesser extent) Aix are exceptional among basal Anatini in the virtual lack of brood attendance by males (Kear 1970). This pattern strongly suggests that biparental brood-rearing is a primitive charac- ter among Anatini (contra Kear 1970), and this polarity explains, at least in part, the prepon- derance of southern-hemispheric dabbling ducks having protracted pair-bonds (Weller 1968, Kear 1970, McKinney 1985). This trait pre- sumably coevolved with a number of characters including migratory habit, seasonal plumage dichromatism, predictability of food resources, and tertiary sex-ratios (Weller 1968, McKinney 1985). Most Anatini (especially Anateae) are char- acterized by monosyllabic distress calls of downy young (Lorenz 1951-1953) and preflight "in- tention" movements (McKinney 1965). A num- ber of presumably derived courtship behaviors further characterize most or all members of the supergenera Lophonetta and Anas. These include marked "inciting displays," variably long "de- crescendo calls," and postcopulatory bathing by females; ritualized head-turning and wing- preening displays by males; and precopulatory "head-pumping" displays by both sexes (Lo- renz 1951-1953; Johnsgard 1961a, 1965). Other behavioral characters, including most "comfort movements" (McKinney 1965), photoperiodic rhythms of reproduction (Murton and Kear 1975, !978; Kear and Murton 1976), and other details of courtship displays (Johnsgard 1961a, 1962, 1965, 1978), are less readily interpreted phylo- genetically, primarily because of inadequate in- formation for a number of species. Reproductive parameters.--Selected quantita- tive parameters of reproduction also show mod- erately pronounced phylogenetic patterns (Ta- ble 1). Hierarchical relationships preclude simple statistical comparisons among members (Felsenstein 1985), but correlations provide use- ful indices to patterns within the tribe. Rohwer (1988) found that the mean egg masses of 53 species of Anatini was positively correlated with mean body masses (r = 0.85). The egg masses have phylogenetic patterns similar to those de- scribed for body mass (Table 1). As noted by Rohwer (1988), insular isolates represent sig- nificant, statistically influential deviations from the overall relationship between egg mass and body mass. Notable examples among Anatini are A. eatoni, A. chlorotis, and (especially) A. auck- landica (Table 1). The evolutionary implications of egg size in A. aucklandica are considered else- where (Livezey 1990). Clutch size (data from Rohwer 1988) is not correlated with body mass among 53 species of Anatini (r = 0.12, log-transformed data). Ex- treme examples include the small clutch-sizes of the comparatively massive S. specularis and A. sparsa, and the large clutch-sizes of tiny Net- tapus (Table 1). This finding, and a modest neg- ative correlation (Rohwer 1988) between clutch size and egg mass in 53 Anatini (r = -0.62) or 43 noninsular Anatini (r = -0.39), suggests that clutch mass (product of egg mass and clutch size) is phylogenetically constrained (see Cody 1966), and that, of the two interrelated param- eters, egg mass has the greater phylogenetic component. The primary determinant of clutch sizes of waterfowl is thought to be energy re- sources available to the female (Lack 1968, Klomp 1970), confounded as well by proximate variations in availability of food (Johnsgard 1973). Other egg characteristics, such as relative thickness of shells and membranes, and pro- portions of soluble and insoluble nitrogen in the shell (Tyler 1964), showed no clear phylo- genetic patterns among Anatini. These correl- ative assessments within tribes, like the studies by Laurila (1988) and Rohwer (1988), are only partly successful in discriminating between phylogenetic constraint and adaptive changes of life-history parameters. Comprehensive as- sessments of phylogenetic constraint must in- corporate more completely resolved phyloge- nies and multivariate techniques that use hierarchically nested designs (Grafen 1989, Funk and Brooks 1990). Evolutionary consistencies and rates of character evolution.--The consistency index of the pro- posed phylogenetic hypothesis is moderately TABLE 2. Summary stafisics for consistency indices of p characters by anatomical-developmental group. Character group p Mean Range Definifive plumage & soft parts Analyzed set 107 0.850 0.250-1.000 Additional a 165 0.903 0.250-1.000 Natal plumage 22 0.777 0.250-1.000 Trachea 21 0.865 0.500-1.000 Skeleton Analyzed set 7 0.809 0.333-1.000 Additional a 10 0.866 0.333-1.000 Includes autapomorphies not in formal analysis; list available. high, compared with other cladistic analyses of avian groups, including an earlier, genus-level analysis of Anseriformes (Livezey 1986). This consistency was achieved despite the compar- atively large number of taxa analyzed and the exclusion of unique autapomorphies. General- ly, character consistencies tend to decrease as the number of taxa analyzed increases (Sander- son and Donoghue 1989). In addition, the four major groups of characters had comparable con- sistency indices (Table 2). It is commonly as- sumed that character suites differ in phyloge- netic "reliability" or informativeness (e.g. Delacour and Mayr 1945), and these judgments may influence the weights characters are given in phylogenetic inference (Bryant 1989). A related issue concerns the rates at which characters evolve, which may in turn determine the phylogenetic and taxonomic level at which the characters are informative in a particular group. Although some inequality of effort ex- panded on character-groups was unavoidable, I believe that the distributions of characters an- alyzed are representative of their relative avail- abilities. Tallies of character changes, grouped by both morphological groups and five major taxonomic levels of the resultant phylogeny (Figs. 2, 3), indicate that character types were unevenly distributed among levels within the tree (Table 3). In particular, characters of defin- itive plumages and soft parts were useful throughout the tree but were especially useful at the lowest levels. Relationships at the level of subtribes, supergenera, genera, and subgen- era were supported disproportionately by state changes in characters of natal plumage and tra- cheal anatomy. Nontracheal skeletal characters, of paramount importance for the establishment TAILE 3. Numbers of character changes, by anatom- ical-developmental group, that support topological divergences at five taxonomic levels. Character group Defini- tive exter- Tra- Skele- Taxonomic level nal Natal cheal tal Tribe and subtribe 7 5 1 5 a Supergenus and genus 41 5 9 3 Cohort and subgenus 21 8 10 0 Infragenus and species-group 76 9 6 1 Species 72+ 1 2 4 Analyzed set 14 1 2 1 Additional b 58 + 0 0 3 Includes four skeletal synapomorphies of Anatinae (Livezey 1986). Characters not included n formal analysis but compiled to confirm monophyly of species; list available on request. of genus-level relationships in Anseriformes (Livezey 1986), were primarily useful in this analysis for resolutions at the tribal and sub- tribal levels (Table 3). This heterogeneity of character change strongly suggests that char- acters evolve at significantly different rates. They are of greatest phylogenetic utility for segments of the tree where rates of evolutionary change are moderately high. Consequently systematists should select those character suites that evo- lutionarily "target" the phylogenetic relation- ships of concern. Osteological studies of Anatini indicate a paucity of phylogenetically informative char- acters of the skeleton (Woolfenden 1961, Live- zey 1986, present study), which underscores the need for extreme caution in the classification of fossil dabbling ducks. Although avian paleon- tologists generally are cognizant of the diffi- culties of classifying fragmentary specimens (e.g. Howard 1964, Olson 1985), the phenetic as- signment of fossils and widespread preoccu- pation with discovery of possibly ancestral lin- eages have seriously compromised the contributions of traditional paleontology to an understanding of avian phylogeny (Cracraft 1980). Based on an osteologically based analysis of Recent genera of Anseriformes (Livezey 1986), many fossil anseriforms, including a number of purported anatines and most pre-Pliocene forms assigned to "Anas," were incorrectly classified (Livezey and Martin 1988, Livezey 1989b). Ten post-Miocene fossils assigned to Anas (or ge- neric synonyms) appear, on the basis of the compendia of Brodkorb (1964) and Howard (1964), to have been classified based on assess- ments of size and phenetic comparisons of un- reliable skeletal features. Size is not phyloge- netically deterministic within the Anatini, and most skeletal variation has not proved reliable for classification. Although analysis of fossils can contribute much to phylogenetic system- atics (Schoch 1986), I agree with Howard (1964: 237) that evolutionary insights to be gained from fossil waterfowl must await an improved knowledge of phylogenetic relationships among Recent representatives of the Anseriformes. ACKNOWLEDGMENTS This research was funded by National Science Foundation grant BSR-8516623. I am indebted to G. Mack, H. Levenson, and R. L. Zusi for their hospi- tality, and to P.S. Humphrey for discussions and encouragements concerning the phylogery of water- fowl. Curatorial staffs of the following institutions permitted access to or loans of study skins, mounted tracheae, or skeletal specimens of waterfowl: Division of Birds, National Museum of Natural History, Wash- ington, D.C. (USNM); Department of Ornithology, American Museum of Natural History, New York (AMNH); Wildfowl Trust, Slimbridge, Gloucester, United Kingdom (WT); Division of Birds, Museum of Zoology, University of Michigan, Ann Arbor; Sub- department of Ornithology, British Museum (Natural History), Tring, Hertfordshire, United Kingdom; Di- vision of Birds, Field Museum of Natural History, Chicago (FMNH); Department of Ornithology, San Diego Museum of Natural History, San Diego; Mu- seum of Natural Science, Louisiana State University, Baton Rouge; Division of Vertebrate Zoology, Bernice P. Bishop Museum, Honolulu; Department of Orni- thology, Royal Ontario Museum, Toronto; and De- partment of Ornithology, Los Angeles County Mu- seum of Natural History, Los Angeles. I thank P.S. Humphrey and R. F. Johnston for access to research facilities at the Museum of Natural History, Univer- sity of Kansas. I appreciate the helpful comments of P.S. Humphrey, and D. G. Homberger, and an anon- ymous referee, and I am grateful for the secretarial assistance of M. Schmalz. LITERATURE CITED AMADON, D. 1966. Insular adaptive radiation among birds. Pp. 18-30 in The Galfipagos: proceedings of the symposia of the Galfipagos International Scientific Project (R. I. Bowman, Ed.). Berkeley, Univ. California Press. ANDERSSON, M. 1986. Evolution of condition-depen- dent sex ornaments and mating preferences: sex- ual selection based on viability differences. Evo- lution 40: 804-816. ANKNEY, C. D., D. G. DENNIS, L. N. WISHARD, & J. E. SEEB. 1986. LOW genic variation between Black Ducks and Mallards. Auk 103: 701-709. --, D. G. DENNIS, & R. O. BAILEY. 1987. Increas- ing Mallards, decreasing American Black Ducks: coincidence or cause and effect? J. Wildl. Manage. 51: 523-529. AWSE, J. C., C. D. ANKNE3/4, & W. S. NELSON. 1990. Mitochondrial gene trees and the evolutionary relationship of Mallard and Black Ducks. Evo- lutionary 44: 1109-1119. BAKER, R. E., & G. A. P^RKER. 1979. The evolution of bird coloration. Philos. Trans. R. Soc. London (B) 287: 63-130. BELLROSE, F. C. 1976. Ducks, geese and swans of North America. Harrisburg, Pennsylvania, Stack- pole Books. --, T. G. SCOTT, A. S. HAWKINS, & J. B. LOW. 1961ß Sex ratios and age ratios in North American ducks. Illinois Nat. Hist. Surv. Bull. 27: 391-474. BOCK, W. J., & J. FARRAND JR. 1980. The number of species and genera of Recent birds: a contribution to comparative systematics. Am. Mus. Novit. No. 2703: 1-29. BOTTJER, P. D. 1983. Systematic relationships among the Anatidae: an immunological study with a his- tory of anatid classification, and a system of clas- sification. Ph.D. dissertation, New Haven, Yale Univ. BREITWISCH, g. 1989. Mortality patterns, sex ratios, and parental investment in monogamous birds. Pp. 1-50 in Current ornithology, vol. 6 (D. M. Power, Ed.). New York, Plenum Press. BREWER, G. 1989. Biparental care behavior of captive Ringed Teal Callonetta leucophrys. Wildfowl 40: 7- 13. BRODKORB, P. 1964. Catalogue of fossil birds: part 2 (Anseriformes through Galliformes). Bull. Flor- ida State Mus. (Biol. Sci.) 8: 195-335. BRODSKY, L. M., & P. J. WEATHERHEAD. 1984. Behav- ioral and ecological factors contributing to Amer- ican Black Duck-Mallard hybridization. J. Wildl. Manage. 48: 846-852. BRUSH, A.H. 1976. Waterfowl feather proteins: anal- ysis of use in taxonomic studies. J. Zool. London 179: 467-498. BRYANT, }-I. N. 1989. An evaluation of cladistic and character analyses as hypothetico-deductive pro- cedures, and the consequences for character weighting. Syst. Zool. 38: 214-227. BUITRON, D., & G. L. NUECHTERLEIN. 1989. Male pa- rental care of Patagonian Crested Ducks Anas (Lophonetta) specularioides. Wildfowl 40: 14-21. BUTCHER, G. S., & S. ROHWER. 1989. The evolution of conspicuous and distinctive coloration for communication in birds. Pp. 51-108 in Current ornithology, vol. 6 (D. M. Power, Ed.). New York, Plenum Press. BUTENDIECK, E., & H. WISSDORF. 1982. Contribution to the designation of the headbones of the Turkey (Meleagris gallopavo) with regard to the NAA (1979). Zool. Jahrb. Anat. 107: 153-184. CODY, M. L. 1966. A general theory of clutch size. Evolution 20: 174-184. COTTER, W. B. JR. 1957. A serological analysis of some anatid classifications. Wilson Bull. 69: 291-300. CRACRAFT, J. 1980. Phylogenetic theory and meth- odology in avian paleontology: a critical apprais- al. Contrib. Sci. Nat. Hist. Mus. Los Angeles County No. 330: 9-16. ß 1983. Species concepts and speciation anal- ysis. Pp. 159-187 in Current ornithology, vol. 1 (R. F. Johnston, Ed.). New York, Plenum Press. ß 1988. Speciation and its ontology: the em- pirical consequences of alternative species con- cepts for understanding patterns and processes of differentiation. Pp. 28-59 in Speciation and its consequences (D. Otte and J. A. Endler, Eds.). Sunderland, Sinauer Assoc. DEL^COUR, J. 1936. Note sur la classification des Anatides. L'Oiseau (n.s.) 6: 366-379. 1954. The waterfowl of the world, vol. i. London, Country Life. --. 1956. The waterfowl of the world, vol. 2. London, Country Life. 1959. The waterfowl of the world, vol. 3. London, Country Life. 1964. Corrections and additions. Pp. 327- 354 in The waterfowl of the world, vol. 4 (J. De- lacour, Ed.). London, Country Life. , & E. M^YR. 1945. The family Anatidae. Wil- son Bull. 57: 3-55. DERSCHEID, J. M. 1938. [Untitled note on systematic positions of Amazonetta and Callonetta.] Bull. Br. Ornithol. Club 58: 62-63. DILGER, W. C., & P. A. JOHNSGARD. 1959. Comments on "species recognition" with special reference to the Wood Duck and the Mandarin Duck. Wil- son Bull. 71: 46-53. DUMBELL, G. 1986. The New Zealand Brown Teal: 1845-1985. Wildfowl 37: 71-87. ELDRIDGE, J. L. 1979. Display inventory of the Tor- rent Duck. Wildfowl 30: 5-15. ENDLER, J.A. 1978. A predator's view of animal color patterns. Pp. 319-364 in Evolutionary biology, vol. 11 (M. K. Hecht, W. C. Steere, and B. Wallace, Eds.). New York, Plenum Press. FELSENSTEIN, J. 1985. Phylogenies and the compar- ative method. Am. Nat. 125: 1-15. FroTH, H.J. 1967. Waterfowl in Australia. Honolulu, East-West Center Press. FUNK, V. A., & D. R. BROOKS. 1990. Phylogenetic systematics as the basis of comparative biology. Smithsonian Contrib. Bot. No. 73: 1-45. GILLESPIE, G.D. 1985. Hybridization, introgression, and morphometric differentiation betwen Mal- lard (Anas platyrhynchos) and Grey Duck (Anas superciliosa) in Otago, New Zealandß Auk 102: 459- 469. GRAVEN, A. 1989. The phylogenetic regression. Phil- os. Trans. R. Soc. London (B) 326: 119-157. HAILMAN, J.P. 1959. Why is the male Wood Duck strikingly colorful? Am. Nat. 93: 383-384. HEINROTH, O. 1911. Beitrige zur Biologie, namen- tlich Ethologie und Psychologie der Anatiden. Pp. 589-702 in Verhandlungen des V. Interna- tionalen Ornithologen-Kongresses (H. Schalow, Ed.). Berlin, Deutsche Ornithologische Gesell- schaft. HEPP, G. R., J. M. NOVAKß K. T. SCRIBNER, & P. W. $TANGEL. 1988. Genetic distance and hybridiza- tion of Black Ducks and Mallards: a morph of a different color? Auk 105: 804-807. HUESMANN, H. W. 1974. Mallard-Black Duck rela- tionships in the Northeast. Wildl. Soc. Bull. 2: 171-177. HILLIS, D.M. 1987. Molecular versus morphological approaches to systematics. Annu. Rev. Ecol. Syst. 18: 23-42. HOSKING, E., & J. KEAR. 1985ß Wildfowlß New York, Facts-on-File Publ. HOWARD, H. 1964. Fossil Anseriformes. Pp. 233-326 in The waterfowl of the world, vol. 4 (J. Delacour, Ed.). Londonß Country Life. HUEY, W. S. 1961. Comparison of female Mallard with female New Mexican Duck. Auk 78: 428- 431. HUMPHREY, P.S., & G. A. CLARK JR. 1964. The anat- omy of waterfowl. Pp. 167-232 in The waterfowl of the world, vol. 4 (J. Delacour, Ed.). London, Country Life. ß & K. C. PARKES. 1959. An approach to the study of molts and plumages. Auk 76: 1-31. HUMPHRIES, C.J. 1983. Primary data in hybrid anal- ysis. Pp. 89-103 in Advances in cladistics, vol. 2 (N. I. Platnick and V. A. Funk, Eds.). New Yorkß Columbia Univ. Press. JACOBß J. 1982. Integumentlipide-ihre chemische Struktur und ihre Bedeutung als systematisches Merkmal in der Zoologie. Funkt. Biol. Med. 1: 83-90. ß & A. GtaSER. 1975. Chemotaxonomy of An- seriformes. Biochem. Syst. Ecol. 2: 215-220. JOHNSGARD, P.A. 1960a. Comparative behavior of the Anatidae and its evolutionary implications. Wildfowl II: 31-45. --. 1960b. Hybridization in the Anatidae and its taxonomic implications. Condor 62: 25-33. --. 1960c. The systematic position of the Ringed Teal. Bull. Br. Ornithol. Club 80: 165-167. --. 1960d. Pair-formation mechanisms in Anas (Anatidae) and related genera. Ibis 102: 616-618. 1960e. A quantitative study of sexual be- havior of Mallards and Black Ducks. Wilson Bull. 72: 133-155. --. 1961a. The taxonomy of the Anatidae--a behavioural analysis. Ibis 103a: 71-85. ß 1961b. Tracheal anatomy of the Anatidae and its taxonomic significance. Wildfowl 12: 58-69. ß 1961c. The systematic position of the Mar- bled Teal. Bull. Br. Ornithol. Club 81: 37-43. ß 1961d. Evolutionary relationships among the North American mallards. Auk 78: 3-43. ß 1962. Evolutionary trends in the behaviour and morphology of the Anatidae. Wildfowl 13: 130-148. 1963. Behavioral isolating mechanisms in the family Anatidae. Pp. 531-543 in Proceedings 13th International Ornithological Congressß vol. I. (C. G. Sibley, Ed.). Washingtonß D.C., Am. Or- nithol. Union. 1965. Handbook of waterfowl behavior. Londonß Constable & Co. ß 1966. The biology and relationships of the Torrent Duck. Wildfowl 17: 66-74. ß 1967. Sympatry changes and hybridization incidence in Mallards and Black Ducks. Am. Midl. Nat. 77: 51-63. 1968a. Some putative Mandarin Duck hy- brids. Bull. Br. Ornithol. Club 88: 140-148. ß 1968b. Waterfowl: their biology and natural history. Lincoln, Univ. Nebraska Press. ß 1973. Proximate and ultimate determinants of clutch size in Anatidae. Wildfowl 24: 144-149. ß 1978. Ducksß geese, and swans of the world. Lincoln, Univ. Nebraska Press. 1979. Order Anseriformes. Pp. 425-506 in Checklist of birds of the world, vol. I, 2nd ed. (E. Mayr and C. W. Cottrell, Eds.). Cambridge, Mas- sachusetts, Mus. Comp. Zool. , & R. DI $ILVESTRO. 1976. Seventy-five years of changes in Mallard-Black Duck ratios in east- ern North America. Am. Birds 30: 905-908. KALTENHAUSER, D. 1971. /ber Evolutionsvorginge in der Schwimmentenbalz. Z. Tierpsychol. 29: 481-540. KEAR, J. 1970. The adaptive radiation of parental care in waterfowl. Pp. 357-392 in Social behaviour in birds and mammals: essays on the social ethology of animals and man (J. H. Crook, Ed.). New Yorkß Academic Press. ß 1972. The Blue Duck of New Zealand. Living Bird 11: 175-192. 1975. Salvadori's Duck of New Guinea. Wildfowl 26: 104-111. ., & R. K. MURTON. 1976. The origins of Aus- tralian waterfowl as indicated by their photore- sponses. Pp. 83-97 in Proceedings of the 16th International Ornithological Congress (H. J. Frith and J. H. Calaby, Eds.). Canberra, Australian Acad. Sci. KESSLER, L. G., & J. C. AVISE. 1984. Systematic rela- tionships among waterfowl (Anatidae) inferred from restriction endonuclease analysis of mito- chondrial DNA. Syst. Zool. 33: 370-380. KLOMP, O. 1970. The determination of clutch-size in birds. Ardea 58: 1-124. KLOPFER, P. 1959ß An analysis of learning in young Anatidae. Ecology 40: 90-102. KURODA, N. 1939. Geese and ducks of the world. Tokyoß publ. privately. [In Japanese.] LACK, D. 1968. The significance of clutch-size in waterfowl. Wildfowl 88: 125-128. 1970. The endemic ducks of remote islands. Wildfowl 21: 5-10. LAURILA, T. 1988. Reproductive strategies in water- fowl: the effect of ultimate environmental fac- tors, size and phylogeny. Ornis Fennica 65: 49- 64. LIVEZEY, B.C. 1986. A phylogenetic analysis of Re- cent anseriform genera using morphological characters. Auk 103: 737-754. 1989a. Phylogenetic relationships and in- cipient flightlessness in the extinct Auckland Is- lands Merganser. Wilson Bull. 101: 410-435. 1989b. Phylogenetic relationships of several subfossil Anseriformes of New Zealand. Univ. Kansas Mus. Nat. Hist. Occas. Pap. 128: 1-25. 1990. Evolutionary morphology of flight- lessness in the Auckland Islands Teal. Condor 92: 639-673. ß & L. D. MARTIN. 1988. The systematic po- sition of the Miocene anatid Anas[?] blanchardi Milne-Edwards. J. Vertabr. Paleontol. 8:196-211. LORENZ, K. 1941. Vergleichende Bewegungstudien an Anatiden. J. Ornithol. 89 (Erg. Bd.): 194-294. ß 1951-1953. Comparative studies on the be- haviour of Anatinae. Avic. Mag. 57: 157-182; 58: 8-17, 61-72, 86-94ß 172-184; 59: 24-34, 80-91. MADGE, $., & H. BURN. 1988. Waterfowl: an identi- fication guide to the ducksß geese and swans of the world. Boston, Houghton Mifflin. MADSEN, C. S., K. P. McHtSGH, & S. R. DE KLOET. 1988. A partial classification of waterfowl (Anatidae) based on single-copy DNA. Auk 105: 452-459. MAYR, E. 1931. Anatomie und systematischen Stel- lung der Salvadori-Ente (Salvadorina waigiuensis Rothsch. & Hartert). Ornithol. Monatsberichte 39: 69-70. ß 1969. Principles of systematic zoology. New York, McGraw-Hill. MCKINNEY, F. 1953. Studies on the behaviour of the Anatidae. Ph.D. dissertation, Bristol, Englandß Univ. Bristol. . 1965. The comfort movements of Anatidae. Behaviour 25: 120-220. ß 1970. Displays of four species of blue-winged ducks. Living Bird 9: 29-64. 1975. The evolution of duck displays. Pp. 331-357 in Function and evolution in behaviour (G. Baerends, C. Beer, and A. Manningß Eds.). Oxford, Clarendon Press. 1985. Primary and secondary male repro- ductive strategies of dabbling ducks. Pp. 68-82 in Avian monogamy (P. A. Gowarty and D. W. Mock, Eds.). Washington, D.C., Am. Ornithol. Union. ß & G. BREWER. 1989. Parental attendance and brood care in four Argentine dabbling ducks. Condor 91: 131-138. ß W. R. SIEGFRIED, I. J. BALLß & P. G. H. FROST. 1978. Behavioral specializations for river life in the African Black Duck (Anas sparsa Eyton). Z. Tierpsychol. 48: 349-400. ', L. G. $ORENSON, r M. HART. 1990. Multiple functions of courtship display in dabbling ducks (Anatini). Auk 107: 188-191. MCKITRICK, M. C., & R. M. ZINK. 1988. Species con- cepts in ornithology. Condor 90: 1-14. MrIKOVSK', J. 1989. Note on the osteology and tax- onomic position of Salvadori's Duck Salvadorina waigiuensis (Aves: Anseridae [Anatidae]). Bull. Br. Ornithol. Club 109: 22-25. MORGANß R. P., II, L. A. NOE, & C. J. HENNY. 1976. Biochemical identification of the Mallard, Anas platyrhynchos, and Black Duckß A. rubripes. Comp. Blochem. Physiol. (B) 53: 499-503. MOULTON, D. W., & M. W. WELLER. 1984. Biology and conservation of the Laysan Duck (Anas lay- sanensis). Condor 86: 105-117. MURTON, R. K., & J. KEAR. 1975ß The role of day- length in regulating the breeding seasons and distribution of wildfowl. Pp. 337-357 in Light as an ecological factor: II (G. C. Evans, R. Bainbridgeß and O. Racham, Eds.). Oxford, Blackwell Sci. Publ. , & --. 1978. Photoperiodism in water- fowl: phasing of breeding cycles and zoogeog- raphy. I. Zool. London 186: 243-283. NEITHAMMER, G. 1952. Zur Anatomie und systema- tischen Stellung der Sturzbach-Ente Merganetta artnata. J. Ornithol. 93: 357-360. NORMAN, F. I., & F. McKINNEY. 1987. Clutches, broods, and brood care behavior in Chestnut Teal. Wildfowl 38: 117-126. NUMACHI, K., g. WATADA, R. KKZWA, N. KURODA, & S. UTIDA. 1983. Evolutionary genetics of the Anatidae. Tori 32: 63-74. OATES, D. W., J. E. $EEB, L. N. WISHARD, r P. ABERSOLD. 1983. Biochemical identification of North Amer- ican waterfowl. Nebraska Game and Parks Comm., Tech. Set. No. 13: 1-68. OrsoN, S.L. 1985. The fossil record of birds. Pp. 79- 238 in Avian biology, vol. 8 (D. S. Farner, J. R. King, and K. C. Parkes, Eds.). New York, Aca- demic Press. ORING, L.W. 1968. Growth, molts, and plumages of the Gadwall. Auk 85: 355-380. PARMER, R. S. (EI.). 1976. Handbook of North Amer- ican birds, vol. 2. New Haven, Connecticut, Yale Univ. Press. PATERSON, H. E.H. 1985. The recognition concept of species. Pp. 21-29 in Species and speciation (E. S. Vrba, Ed.). Pretoria, Transvaal Museum. 1988. On defining species in terms of ste- rility: problems and alternatives. Pacific Sci. 42: 65-71. PAtrON, J. C., & J. C. AVlSE. 1985. Evolutionary ge- netics of birds IV: rates of protein divergence in waterfowl (Anatidae). Genetica 68: 129-143. PETERS, J.L. 1931. Check-list of birds of the worldß vol. 1. Cambridge, Harvard Univ. Press. PHILLIPS, J.C. 1912. A reconsideration of the Amer- ican Black Duck with specific reference to certain variations. Auk 29: 295-306. . 1922. A natural history of the ducks, vol. 1. Boston, Houghton Mifflin. 1923. A natural history of the ducks, vol. 2. Boston, Houghton Mifflin. 1925. A natural history of the ducks, vol. 3. Boston, Houghton Mifflin. 1926. A natural history of the ducksß vol. 4. Bostonß Houghton Mifflin. RIPLEY, S. D. 1942. A review of the species Anas castanea. Auk 59: 90-99. 1951. Remarks on the Philippine Mallard. Wilson Bull. 63: 189-191. 1957a. A paddling of ducks. New Yorkß Har- court, Brace and Co. . 1957b. The waterfowl of the world [review of Delacour 1956]. Auk 74: 269-272. ROHWER, F. C. 1988. Inter- and intraspecific rela- tionships between egg size and clutch size in waterfowl. Auk 105: 161-176. SALVADORI, T. 1895. Catalogue of the Chenomor- phae (Palamedeae, Phoenicopteri, Anseres), Crypturi and Ratitae in the collection of the Brit- ish Museum. London, Longroans. SANDERSON, M. J., & M. J. DONOGHUE. 1989. Patterns of variation in levels of homoplasy. Evolution 43: 1781-1795. SCHERER, S., & T. HILSBERG. 1982. Hybridisierung und Verwandtschaftsgrade innerhalb der Anat- idae--eine systematische und evolutionstheore- tische Betrachtung. J. Ornithol. 123: 357-380. ß & C. SONTAG. 1986. Zur molekularen Tax- onomie und Evolution der Anatidae. Z. Zool. Syst. Evolut.-forsch. 24: 1-19. ScHocH, R. M. 1986. Phylogeny reconstruction in paleontology. New Yorkß Van Nostrand Rein- hold. SCHVTZ, F. 1965. Sexuelle Pragung bei Anatiden. Z. Tierpsychol. 22: 50-103. SELANDER, g. K. 1972. Sexual selection and dimor- phisin in birds. Pp. 180-230 in Sexual selection and the descent of man: 1871-1971 (B. G. Camp- bell, Ed.). Chicago, Aldine Publ. SIELEY, C.G. 1957. The evolutionary and taxonomic significance of sexual dimorphism and hybrid- ization in birds. Condor 58: 166-191. -, & J. E. AHLQVIST. 1972. A comparative study of the egg-white proteins of non-passerine birds. Peabody Mus. Nat. Hist. Bull. 39: 1-276. SIEGFRIED, W. R. 1974. Brood care, pair bonds and plumage in southern African Anatini. Wildfowl 25: 33-40. SOOTHILL, E., & P. WmHD. 1978. Wildfowl of the world. Poole, Dorset, Blandford Press. STAHL, J.-C., J.-L. MOrGaN, P. JOVVENtN, & H. WE- MERSKIRCH. 1984. Le canard d'Eaton, Anas eatoni drygalskii, des Iles Crozet: systematique, com- portement alimentaire et biologie de reproduc- tion. Gerfaut 74: 305-326. SWOIEORl, D.L. 1985. PAUP: phylogenetic analysis using parismony. Version 2.3. Urbana, Illinois Nat. Hist. Surv. TolI, F.S. 1979. Waterfowl: ducks, geese and swans of the world. New York, Harcourt-Brace Jova- novich. TPaVERS, R.L. 1972. Parental investment and sexual selection. Pp. 136-179 in Sexual selection and the descent of man: 1871-1971 (B. G. Campbellß Ed.). Chicago, Aldine Publ. TYLER, C. 1964. A study of egg shells of the Anatidae. Proc. Zool. Soc. London 142: 547-583. VERHEYEN, R. 1955. La systmatique des Ansri- formes base sur l'ostologie compare. Bull. Inst. R. Soc. Nat. Belgique 31(37): 1-22, (38): 1-16. VON BOETTICHER, H. 1942. Ober die Einteilung der Familie der Entenv6gel (Anatidae) in Unterfa- milien und Sektionen. Zool. Anz. 140: 37-48. --. 1952. Gnse- und Entenv6gel aus allerWelt. Leipzig, Germany, Geest & Portig K.-G. WAGNER, W. H. JR. 1983. Reticulistics: the recogni- tion of hybrids and their role in cladistics and classification. Pp. 63-79 in Advances in cladistics, vol. 2 (N. I. Platnick and V. A. Funk, Eds.). New York, Columbia Univ. Press. WARNER, g.W. 1971. The structural basis of the or- gan of voice in the genera Anas and Aythya (Aves). J. Zool. London 164: 197-207. WELLER, M.W. 1964. Distribution and species rela- tionships. Pp. 108-120 in The waterfowl of the world, vol. 4 (J. Delacour, Ed.). London, Country Life. --. 1968. Notes on some Argentine anatids. Wil- son Bull. 80: 189-212. --. 1980. The island waterfowl. Ames, IowaState Univ. Press. WEsT-EBERHARD, M.J. 1983. Sexual selection, social competition, and speciation. Q. Rev. Biol. 58: 155- 183. WIL, E. O. 1981. Phylogenetics: the theory and practice of phylogenetic systematics. New Yorkß J. Wiley & Sons. WILLIAMS, D.M. 1983. Mate choice in the Mallard. Pp. 297-309 in Mate choice (P. Bateson, Ed.). Cam- bridge, Cambridge Univ. Press. WOLtERS, H.E. 1976. Die Vogelarten der Erde, part 1. Hamburg and Berlin, Paul Parey. WOOLFErriES, G. E. 1961. Postcranial osteology of the waterfowl. Florida State Mus. Bull. 6: 1-129. YAMASHINA, Y. 1948. Notes on the Marianas Mal- lard. Pacific Sci. 2: 121-124. 1952. Classification of the Anatidae based on cyto-genetics. Pap. Coord. Comm. Res. Gen. 3: 1-24. APPENDIX I. Characters used in the phylogenetic analysis of Anatini. The sequence corresponds to that used in the figures and text. Presumed primitive character states are defined as "a"; derived states are designated alphabetically thereafter (with the exception of character 27). Characters with two or more derived states, with the exception of rectrix number (character 27), were analyzed as unordered. Characters are grouped by general anatomical or developmental class, but groupings of some bill characters that involved the integument and underlying bones were necessarily arbitrary. To avoid biased comparisons of character groups, two particularly difficult cases involving bill shape (characters 97 and 157) were assigned to "soft- parts" and "skeleton," respectively. Potentially confusable characters are cross-referenced parenthetically. Unless indicated otherwise, characters of definitive plumages pertain to both alternate and basic plumages and to both sexes. Characters of syringeal bulla pertain to males. Consistency indices (CI) for each character correspond to the topology given in Figures 2 and 3. ADULT (DFINITIVE) PLUMAGES 1. Tail: (a) moderately short; (b) long, broad, rounded, generally black (extends well beyond feet in normally prepared skins); CI = 0.50. 2. Rump of definitive males: (a) not as follows; (b) with contrasting black lateral margins; CI = 0.50. 3. Lateral vanes of outer rectrlces: (a) colored as medial vanes; (b) whitish, contrasting with medial vanes, producing fine pale borders of tail; CI = 0.50. 4. Central pair of rectrices: (a) of approximately equal length to other rectrices; (b) distinctly elongate relative to other rectrices; CI = 1.00. 5. Central pair of rectrices: (a) straight; (b) dorsally curled; CI = 1.00. 6. White vertical shoulder mark: (a) absent; (b) present in males; CI = 0.50. 7. Narrow white antorbital facial crescent: (a) absent; (b) present in males (only in basic of A. clypeata; occurs also in hybrids involving A. clypeata, see Scherer and Hilsberg 1982); CI  1.00. 8. Pale flank patches, formed by laterodorsal extensions of pale venter to sides of rump: (a) absent; (b) present in males, alternate plum- age (compare with character 18); CI = 0.33. 9. Secondary coverts pale blue, forming contrasting blue forewing patch: (a) absent; (b) present (excludes more extensive, gray forewing of A. querquedula and infregenus Punanetta); CI  1.00. 10. Distinct, whitish "horseshoe"-shaped marks on breast feathers, all definitive plumages (exclusive of alternate plumage of male A. cly- peata): (a) absent; (b) present (variable, weak in A. smithii; compare with character 38); CI = 0.50. 11. Dark, contrasting cap, extending to below orbit and including nape: (a) absent; (b) present; CI = 0.50. 12. Contrasting, white vertical neck stripe: (a) absent; (b) present (reduced in A. eatoni; Stahl et al. 1984); CI = 1.00. 13. Secondary coverts white, forming contrastingly pale forewing patch: (a) absent; (b) present; CI  1.00. 14. Contrasting, pale buff or white forecrown: (a) absent; (b) present; CI = 1.00. 15. Greenish postorbital patch, washed with purplish-bronze iri- descence, lacking contrasting border: (a) absent; (b) present (compare with characters 28, 31); CI  1.00. 16. Smudgy black crown, from bill base to occiput: (a) absent; (b) present in definitive alternate plumages of males; CI = 1.00. 17. Sharply defined, narrow white eye-ring: (a) absent; (b) present, both sexes; CI  1.00. 18. Black-bordered buff flank patches: (a) absent; (b) present in males, alternate plumage; CI = 0.50. 19. Rump: (a) not as follows; (b) dark brown with greenish irides- cence, in males; CI = 1.00. 20. Broad white stripes on lateral vanes of tertials: (a) absent; (b) present in both sexes; CI = 1.00. 21. "True" speculum, a contrastingly colored, metallic patch limited to secondaries: (a) absent (includes the nonmetallic white "speculum" or generalized alar iridescence of Cairineae and Chenonetteae); (b) present; CI = 1.00. 22. Color of (true) speculum (where present; not codable for autapo- morphic M. strepera): (a) metallic purplish-bronze; (b) metallic green and blue, variably washed with purplish or black (obsolete in A. ber- nieri); CI  1.00. 23. Rufous or pinkish-red sides: (a) absent; (b) present, both sexes; CI = 1.00. 24. Pinkish-brown wash on breast: (a) absent; (b) present, both sexes; CI  1.00. 25. Cinnamon wash on sides and belly: (a) absent; (b) present in males, alternate plumage; CI = 1.00. 26. Chestnut wash on sides, belly and breast: (a) absent; (b) present; CI = 1.00. 27. Number of pairs of rectrices (state "b" considered primitive; modal condition in nonmolting adults): (a) 6; (b) 7; (c) 8; (d) 9; CI = 0.33. 28. Entire head and neck (including chin and throat) of male in alternate plumage a bright, metallic green: (a) not so (includes green iridescence on sides of head in male A. clypeata); (b) as described (ves- tigial in A. oustaleti and A. wybilliana); CI = 0.50. 29. Brick-red coloring of breast, as distinct from sides, belly: (a) absent; (b) present in males, especially in alternate plumage (vestigial in A. fulvigula, A. oustaleti, and A. wyvilliana); CI = 0.50. 30. Generalized, poorly differentiated greenish postorbital irides- cence: (a) absent; (b) present; CI = 1.00. 31. Intense, metallic-green postorbital patch with black and/or buff border: (a) absent; (b) present; CI = 1.00. 32. Entire rump contrastingly black with green iridescence: (a) not so (includes the purplish-brown rump of A. poecilorhyncha and greenish- brown rump of male A. clypeata); (b) as described (vestigial in A. oustaleti, A. wyvilliana, A. laysanensis); CI = 1.00. 33. Sharp, black-and-white vermiculations on undertail coverts: (a) absent; (b) present; CI = 1.00. 34. Sharp, black-and-white barring on flanks: (a) absent; (b) present; CI = 0.50. 35. Fine blackish vermiculations on flanks: (a) absent; (b) present; CI = 1.00. 36. Fine blackish vermiculations on whitish or pale buff belly feath- ers of males: (a) absent; (b) present; CI = 0.50. 37. Constrasting black wing linings: (a) absent; (b) present; CI = 0.50. 38. Alternating brown and white, U-shaped marks on breast feath- ers: (a) absent; (b) present, males in alternate plumage (Fig. 4a; compare with character 10); CI = 0.67. 39. Contrastingly pale caudal border on speculum (where present): (a) present, white or buffy; (b) obsolete; CI = 0.50. 40. Pale caudal border of speculum (where present): (a) narrow, (cranio-caudal) width significantly less than that of metallic speculum; (b) broad, width approximating or exceeding that of speculum; CI = 1.00. 41. Breast feathers of males in alternate plumage: (a) not as follows; (b) with dark central subterminal spot subtended proximally by trans- verse row of four lighter spots (Fig. 4b); CI = 1.00. 42. Cranial border of speculum, formed by tips of caudal-most row of greater secondary coverts: (a) not contrastingly colored; (b) black; (c) white; (d) buff or rufous; CI = 0.38. 43. Tips of greater secondary coverts (character 42) subtended with contrasting white or light brown subterminal band: (a) not so; (b) as described; CI = 1.00. 44. Basal half of greater secondary coverts: (a) brown; (b) white; (c) black; CI = 0.67. 45. Lateral vane of tertials of males in alternate plumage: (a) not as follows; (b) with sharply defined, narrow white edge, bordered me- dially by black (excludes white edge subtended by brown in M. capensis, A. querquedula); CI = 1.00. 46. Scapulars of males in alternate plumage: (a) not as follows; (b) with distally widened, sharply defined, typically asymmetrical, white central stripe (Fig. 5d); CI = 0.50. 47. Scapulars of males in alternate plumage: (a) not as follows; (b) A C t I Fig. 4. Breast feathers of four species of Anateae (definitive alternate plumages of males, ventral views): (A) Mareca falcata (AMNH 732,314), (B) Anas crecca (AMNH 424,195), (C) A. discors (AMNH 749,780), and (D) A. erythrorhyncha (AMNH 262,156). with distal portion of lateral vane blue, and medial vane blackish with distally widening buff stripe along rachis (Fig. 5c); CI = 1.00. 48. Back feathers of both sexes in alternate plumage: (a) brownish with margins variously lighter; (b) finely vermiculated with buff; CI = 1.00. 49. Back feathers of males in alternate plumage: (a) not as follows; (b) crossed by 3-4 straight bars of white; CI = 0.50. 50. Back feathers of females: (a) not as follows; (b) marked in mid- vane by wide, whitish, U-shaped bar; CI = 1.00. 51. Back feathers: (a) not as follows: (b) dichromatic--with buffy- cinnamon crescent distally and similarly colored, irregular line prox- imally in males, and a single central buffy crescent in females; CI = 1.00. 52. Shape of scapulars of males in alternate plumage: (a) typically pinnate; (b) tapering abruptly to long narrow point (Fig. 5c, d); CI = 1.00. 53. Buff mottling on basal half of lateral vanes of scapulars of males in alternate plumage: (a) absent; (b) present; CI = 1.00. 54. Uppertail coverts of both sexes: (a) not as follows; (b) largely black with distal edges of medial vanes buff; CI = 1.00. 55. Scapulars of males in alternate plumage: (a) not as follows; (b) finely vermiculated with gray and white, especially medial vane; CI = 1.00. 56. Sharply defined black chin patch: (a) absent; (b) present in males in alternate plumage; CI = 1.00. 57. Fine, black, circular speckling of head: (a) absent; (b) present on whitish background in basic plumage of males and all definitive plum- ages of females; (c) present on brownish background (all definitive plumages except alternate of male A. acuta); CI = 1.00. 58. Short mane of feathers on nape, relatively well developed in lower (nuchal) portion: (a) absent; (b) present (compare with characters 87, 88); CI = 1.00. 59. Dorsal contour feathers characterized by an irregular buff mar- ginal and one irregular central buff mark: (a) not so; (b) as described; CI = 1.00. 60. One or two V-shaped white marks on dorsal contour feathers: (a) absent; (b) present; CI = 1.00. 61. Entire back and crown iridescent green: (a) not so; (b) as de- scribed; CI = 1.00. 62. Faces of females with three adjacent cheek stripes, alternating white, black, and white: (a) not so; (b) as described; CI = 1.00. 63. Entire underwing, excluding remiges; (a) variably mottled or black; (b) immaculate white; (c) dark chestnut; CI = 1.00. 64. Coloration of body, including dorsum, tail, sides, venter, head, and neck: (a) various, but not as follows; (b) black; (c) dark brown; CI = 1.00. 65. Primary remiges: (a) uniformly dark; (b) dark with fine, whitish lateral margins; CI = 1.00. 66. Alternating black and white crescents on sides of breast: (a) absent; (b) present in males in alternate plumage; C1 = 1.00. 67. Black-and-white scalloping on sides: (a) absent; (b) present in males in alternate plumage; CI = f.00. 68. White, drop-shaped postocular streaks: (a) absent; (b) present in females; CI = 1.00. 69. Crown and nape darkened by a contrastingly dark, dull-brown "hood": (a) not so; (b) present in males in alternate plumage; CI  1.00. 70. Speculum (where present and "true"): (a) not as follows; (b) strongiy washed with iridescent bronze and distinctly dichromatic, those of females having little or obsolete iridescence; CI  1.00. 71. Wings rounded, primary remex I0 distinctly shorter than pri- mary remiges 8 and 9: (a) not so (includes uniquely rounded condition in A. aucklandica, in which both remex 9 and I0 are shorter than remex 8); (b) characteristic of both sexes; CI = 1.00. 72. Rectrices progressively shortened laterally, producing wedge- shaped tail: (a) not so; (b) as described; CI = 1.00. 73. Scapulars of males in alternate plumage: (a) not as follows; (b) patterned with fine, white, obliquely oriented vermiculations (Figs. 5a vs. 5b); CI = 1.00. 74. Entire upper forewing, including all secondary coverts and less- er primary coverts (alular region): (a) not contrastingiy patterned; (b) immaculate white; (c) pale blue; CI = 1.00. 75. Color of lower breast and belly: (a) variously colored, often mottled, but if pale not distinctly demarcated from color of sides, flanks, and breast; (b) white, sharply contrasting with darker sides, fianks, and breast; CI = 1.00. 76. Scapulars of adults, especially males: (a) not as follows; (b) with contrasting black stripe along rachis; CI = 1.00. 77. Uppertail coverts distinctly pointed with pale lateral edges: (a) not so; (b) as described; CI = 1.00. 78. Axillars of adults (both sexes): (a) not as follows; (b) white with minute grayish spotting distally; (c) with minute brown spotting dis- tally; CI = 0.67. 79. Diffuse greenish iridescence of dorsum of wing, if present: (a) variable in intensity, typically including secondary converts, secondary remiges, tertials, and (in some) primary remiges; (b) limited to specular region of secondary remiges; CI = 1.00. 80. Upperwing: (a) not as follows; (b) generally iridescent green (particularly intense on secondary remiges) with laterally widening, white band formed by tips of greater secondary coverts (reduced prox- imafly in N. auritus); (c) generally iridescent green with white band on secondary coverts obsolete; CI = 1.00. 81. Rump paler than rectrices, finely scalloped with dark gray: (a) not so; (b) as described; CI = 1.00. 82. Flanks of males in alternate plumage: (a) not as follows; (b) marked by subrectangular whitish patch surrounded by dark areas; CI 83. Flanks of males in alternate plumage: (a) not as follows; (b) marked by subcircular whitish patch surrounded by dark areas; CI = 1.00. 84. Breast feathers of males in alternate plumage: (a) not as follows; rust, white A brownish B C D I I )lue Fig. 5. Scapulars of four species of Anateae (definitive alternate plumages of males, dorsal views, right pterylae): (A) Mareca americana (AMNH 350,657), (B) M. sibilatrix (AMNH 443, 631), (C) Anas cyanoptera (AMNH 731,952), and (D) A. rhynchotis (AMNH 732,785). (b) with one dark, central terminal spot, subtended by a staggered row of three dark spots, followed by a more-proximal row of four paler spots (Fig. 4c); CI = 1.00. 85. Breast feathers of males in alternate plumage: (a) not as follows; (b) with one dark, central terminal spot, subtended by two variably distinct dark spots, followed by a blurred basal line (Fig. 4d); CI = 0.50. 86. Edge of lateral vane of tertials of males in alternate plumage: (a) not as follows; (b) with sharp defined, narrow black border, sub- tended medially by buffy brown (interrupted proximally by white in A. crecca); CI = 1.00. 87. Long crest, emerging from crown to nape, the longest plumes originating anteriorly: (a) absent; (b) present in both sexes; CI = 1.00. 88. Nuchal crest of moderate, relatively uniform length: (a) absent; (b) present, longer in males (comparatively long in M. falcata; cf. char- acters 58, 87); CI = 0.33. 89. Heads of adults of both sexes: (a) poorly demarcated from breast, or differs qualitatively from breast in color; (b) sharply demarcated quan- titatively from breast in the darkness of brownish-gray color; CI = 1.00. 90. Ground color of axillars (both sexes): (a) white; (b) dark brown to black; CI = 0.25. 91. Axillars: (a) without conspicuous dark transverse barring or spotting; (b) so marked, especially pronounced in females (excludes dull brownish mottling of A. laysanensis and A. aucklandica; includes reduced, subspecifically variable markings of A. bahamensis and almost completely darkened A. georgica; CI = 0.33. So Ps 92- Background color of bill: (a) dark gray; (b) greenish yellow to orange (less bright in A, wyvilliana, A. laysanensis; excludes lemon-yellow of N. auritus); (c) bright bluish gray; (d) scarlet, in males (excludes unique red-colored bills of M. capensis, Amazonetta); CI = 1.00. 93. Lateronasal regions of bill: (a) of same color as rest of bill; (b) marked by contrasting yellow patches; (c) marked by contrasting red patches; CI = 0.50. 94. Sharply defined, dark stripes along culmen, especially conspic- uous in males: (a) absent; (b) present (Fig. 6c; brownish in A. erythrorhyn- cha, obscured by background color in A. bahamensis); CI = 0.50. 95. Dark stripe along tomium and forming variably broad, dark line of base of bill: (a) absent; (b) present; CI = 0.50. 96. Color of feet: (a) dark gray; (b) yellow to reddish orange; (c) bright pink or coral red; CI = 0.29. 97. Bill (lateral view): (a) with concave dorsal ridge, nostrils basal to midpoint of culmen; (b) short, stout, with convex dorsal ridge, nos- trils located approximately at midpoint of culmen (Fig. 6b); CI = 1.00. 98. Base of bill swelling in breeding males: (a) not so; (b) as described (Fig. 6a); CI - 1.00. 99. Variable, flesh-colored subterminal patch on dorsum of bill: (a) absent; (b) present (Fig. 6a); CI = 0.50. 100. Contrasting, dark nasal patches on dorsum of bill (obscured in most older skin specimens): (a) absent; (b) present in both sexes (see photograph in Hosking and Kear [1985: 98]; excludes the perinasal patches of S. specularis and the sexually dichromatic, irregularly shaped, comparatively blurred markings in some members of subgenus Arias); CI = 1.00. 101. Base of bill: (a) with marked angular dorsal prominences (Fig. 6a); (b) sloping, without angular prominence (Fig. 6: b-c); CI - 0.50. 102. Claws of pedal phalanges: (a) straight or moderately curved; (b) strongly curved, very robust; CI = 1.00. 103. Contrastingly dark tomial stripe (without basal extension): (a) absent; (b) present (Fig. 6c); CI = 1.00. 104. Iris color: (a) brown, both sexes; (b) yellow in males, brown in females; (c) reddish-yellow, both sexes; (d) red, both sexes, males bright- er (excludes sexually dichromatic, scarlet irides of Aix sponsa, and vari- able iris color of M. capensis, L. specularioides); CI = 0.75. 105. Cartilaginous carpal knobs (without elongation of underlying process of metacarpal I as in Tadominae): (a) absent; (b) present, often naked of feathers in adult males (excludes suggestions of knobs present in males of some Anateae, e.g.L. specularioides, S. specularis, and A. sparsa); CI = 1.00. 106. Bright yellow patch on upper bill tip, immediately caudad to nail: (a) absent; (b) present (both sexes); CI = 1.00. NATAL PLUMAGES (The most variable and problematical characters used in this analysis are natal characters. Evaluations of states requires reference to series of specimens of very young ducklings. Caution is required to avoid badly "foxed" or faded skins or specimens of uncertain species identification, including hybrid birds from avicultural holdings; detection of facial and dorsal markings can be difficult in the convergently darkened, variable natal plumages of some insular forms, e.g.A. albogularis and A. laysanensis.) 107. Sharp, dark postorbital stripe without any trace of preorbital stripe: (a) absent; (b) present; CI = 0.50. 108. Diffuse, darkish, suborbital cheek patch extending from base of bill: (a) absent; (b) present (Fig. 7b); CI = 0.50. 109. Buffy supraorbital spot: (a) absent; (b) present; CI = 1.00. 110. Background color of face: (a) white or yellowish; (b) washed with tannish-rust; CI = 1.00. D Fig. 6. Bills of four species of Anatini (adult males in breeding season, lateral views): (A) Cairina rnoschata (KUMNH 28,524), (B) Nettapus auritus (KUMNH 36,230), (C) Anas acura (KUMNH 9,593), and (D) A. clypeata (KUMNH 64,924). D Fig. 7. Heads of downy young of six species of Anatini (age class I, lateral views): (A) Nettapus coroman- delianus, female (FMNH 4,323); (B) Mareca capensis, unsexed (AMNH 348,435); (C) Anas fulvigula, male (AMNH 816,980); (D) A. castanea, unsexed (WT 680); (E) A. acura, female (NMNH 299,286); and (F) A. querquedula, male (WT 622). D Fig. 8. Syringeal bullae of eight species of Anatini (adult males, ventral views): (A) Cairina moschata (WT 1,252), (B) Callonetta leucophrys (WT 1,652), (C) Mareca penelope (WT 1,043), (D) Arias rubripes (WT 2,164), (E) A. platalea (WT 922), (F) A. erythrorhyncha (WT 938), (G) A. formosa (WT 130), and (H) A. puna (WT 1,442). 111. Pale scapular and rump spots: (a) rounded, well separated; (b) elongated, the lateral pairs sometimes joining to form pale dorsal stripes; CI = 0.50. 112. Sides, flanks, and belly: (a) colored like breast and throat; (b) distinctly mottled with irregular blotches of dark gray; CI = 1.00. 113. Fine, dark, sharply contrasting breast band: (a) absent; (b) pres- ent (Fig. 7d); CI  0.50. 114. Dark malar spot: (a) absent; (b) present; CI = 1.00. 115. Black-bordered, contrasting orange-buff gape spot: (a) absent; (b) present; CI = 1.00. 116. Darkish orbital stripe passing from bill to nape, sharply defined along dorsal edge but blending with color of cheeks ventrally: (a) absent; (b) present; CI  1.00. 117. Dark auricular spot: (a) absent (Fig. 7a); (b) present (Fig. 7: b-f); CI = 0.50. 1 IS. Cheek stripe (if present) wide, incorporating dark auricular patch, forming the ventral border to a very narrow, pale suborbital stripe, and typically converging broadly with dark orbital stripe anterior to orbit: (a) not so (includes unique two-toned, "false" cheek stripe of Cheno- netta); (b) as described (Fig. 7d); CI = 1.00. 119. Cheek stripe (if present) narrow, well separated from orbital stripe by wide white band (rarely occluded anteriorly): (a) not so (in- cludes unique two-toned cheek stripe of Chenonetta); (b) as described (Fig. 7f); CI = 1.00. 120. Cheek stripe (if present) broad, bordered dorsally by relatively wide white suborbital band, variably convergent with orbital stripe anteriorly, and widening posteriorly: (a) not so (includes unique two- toned, "false" cheek stripe of Chenonetta); (b) as described (Fig. 7c); CI = 1.00. 121. Bright yellow wash on throat and upper breast: (a) absent; (b) present (especially variable in some species); CI = 0.50. 122. Pale stripe between dark orbital and cheek stripes buffy, con- trasting with whitish of throat: (a) absent; (b) present (Fig. 7d); CI = 1.00. 123. Pale supraorbital stripes: (a) terminate at nape (Fig. 7: b-f); (b) continue caudally to meet across nape (Fig. 7a); CI = 1.00. 124. Rectrices: (a) short, not stiflened; (b) disproportionately long, stiff; CI = 1.00. 125. Wing linings: (a) variably mottled with dark; (b) immaculate white or yellow; CI = 0.25. 126. Dark cheek stripe: (a) absent (Fig. 7: a-c); (b) present (Fig. 7: d- f); CI = 0.33. 127. Uneven, variable dark streak passing obliquely from orbit through auricular region to nape: (a) absent; (b) present; CI = 0.50. 128. Dark, variably extensive postorbital stripe (often retained in juvenal and definitive basic plumages of both sexes, as well as definitive basic of males): (a) absent; (b) present (Fig. 7); CI = 1.00. TRACHEA AND SYPaUx (ADULT M^LrS) 129. Medial margin of syringeal bulla: (a) subcircular in outline (Fig. 8: b-c; includes intermediate conditions of M. strepera and M. americana); (b) flattened or concave, bulla often tending toward subtriangular shape (Fig. 8: d-h); C1 = 0.50. 130. Medial extremity of syringeal bulla: (a) variably rounded; (b) distinctly pointed (Fig. 8f; not codable for uniquely conformed bullae of the subgenera Spatula and Punanetta, or for A. querquedula and A. formosa); CI = 0.50. 131. Trachea: (a) without distinct swellings craniad to syrinx; (b) with one (versicolor and puna) or two (hottentota) swellings; CI = 1.00. 132. Syrigneal bulla (uniquely derived bulla of A. formosa not com- parable): (a) large, lateromedial width more than twice that of trachea; (b) small, lateromedial width less than twice that of trachea (Fig. Be, see also Fig. 8g); CI = 0.50. 133. Syringeal bulla: (a) distinctly asymmetrical, medial extremity uninflated (Fig. 8: a-d, f-h); (b) almost symmetrical, medial extremity inflated, producing a distinct medial lobe aligned with trachea (Fig. Be); CI = 0.50. 134. Laterocaudal margin of syringeal bulla: (a) smoothly inflated or slightly depressed (Fig. 8: b-d); (b) with distinct depression in ventral surface (Fig. 8f; not codable for the uniquely conformed bullae of A. querquedula, A. formosa, and the subgenus Punanetta); CI = 1.00. 135. Syringeal bulla uniformly inflated so as to obscure 1obing and produce a subcubic ventral appearance: (a) not so (Fig. 8: a-g); (b) as described (Fig. 8h); CI = 1.00. 136. Medial lobe of syringeal bulla enlarged (at least one-half as large as lateral lobe); ventrally knobbed, and demarcated from lateral lobe by distinct constriction: (a) not so; (b) as described (Fig. 8d); CI = 1.00. 137. Syringeal bulla: (a) absent, conspicuous; (b) obsolete; CI = 1.00. 138. Syringeal bulla: (a) comparatively densely walled, opaque; (b) thinly walled, uniformly translucent; CI = 1.00. 139. Syringeal bulla: (a) rounded ventrally; (b) distinctly dorsoven- trally compressed (Fig. 8a); CI = 1.00. 140. Syringeal bulla: (a) tracheal involvement in syrinx compara- tively small, typically four rings fused into syrinx, and pessulus thin, weakly supported dorsally (Fig. 8: a-b); (b) tracheal chamber large, typically incorporating at least six rings, and pessulus dense, folded, and stoutly supported dorsally (Fig. 8: c-h); CI = 1.00. 141. Distinct ventro-caudal bulge in bulla in vicinity of divergence of bronchii: (a) absent (Fig. 8: a-f); (b) present (Fig. 8: g-h); CI = 0.50. 142. Paired, laterally and medially directed flanges on caudal margin of pessulus (visible within syringeal bulla in caudal view): (a) absent; (b) present; CI = 1.00. 143. Laterocranial elongation of lateral lobe of syringeal bulla: (a) absent; (b) apparent; CI = 1.00. 144. A distinct cranial displacement of lateral lobe of syringeal bulla, producing an indentation lateral to caudal tracheal aperture: (a) absent; (b) present (Fig. 8d); CI = 1.00. 145. Craniocaudal compression of lateral lobe of syringeal bulla, pro- ducing lateral lobe of approximately uniform width: (a) absent; (b) present; CI = 1.00. 146. Medial rim of trachea, within syringeal bulla: (a) remains dis- tinct ventrally; (b) reduced, merging with wall of bulla ventrally; CI = 0.67. 147. Prominent, square, lateral flange of passullus: (a) absent; (b) present; CI = 1.00. 148. Caudodorsal extremity of caudal tracheal aperture of syringeal bulla: (a) rounded; (b) pointed, angular; CI  1.00. 149. Lateral lobe of syringeal bulla: (a) uniformly walled; (b) typically with unossified gap caudally (Fig. 8b); CI = 1.00. SKELETON (EXCLUSIVE OF TRACHEA) 150. Ventral manubrium of sternum: (a) without conspicuous prom- inence; (b) with prominent, peg-like, medial spine (includes L. specu- larioides, state incorrectly given in Livezey 1986: fig. 4); CI = 0.33. 151. Dorsal manubrium of sternum: (a) marked medially by rounded notch; (b) marked medially by a rounded notch enclosing a small, central prominence (variable); CI = 1.00. 152. Distal elongation of entepicondylar process, relative to ectepi- condylar process of humerus: (a) lacking; (b) evident; CI = 1.00. 153. Supraorbital processes of skull: (a) lacking or elongate, essen- tially coplanar with dorsal surface of interorbital surface; (b) substantial, flat, medially appressed to dorsal margin of orbit; (c) moderately elon- gated, with comparatively great dorsolateral prominence; CI = 1.00. 154. Brachial tuberosity of coracold: (a) nonpneumatic; (b) with pneu- matic depression under caudal edge (variable, especially in A. galericu- lata); CI = 0.33. 155. Ventral surface of basioccipital shield of skull: (a) rounded; (b) with medial ridge; CI = 1.00. 156. Posterior surface of orbit, anterolateral view of skull: (a) foramen nervus maxillomandibularis distinctly smaller than foramen opticum; (b) foramen nervus maxillomandibularis approaching size of foramen opticurn; CI = 1.00. 157. Bill shape: (a) typically anatine; (b) spatulate, with anterior third laterally expanded and (externally) lamellae lengthened so as to pro- trude ventrally from tomium (Fig. 6d); CI = 1.00. APPENDIX 2. A Linnean classification of modern dabbling ducks, Tribe Anatini; subtribal names were formed in the manner of the sections given by Boetticher (1942) based on the included genus first elevated to suprageneric rank. Sedis mutabilis indicates that included taxa of next-lower rank are of undetermined sequence. Asterisks mark extinct forms. Currently recognized subspecific taxa included within "polytypic" species are listed parenthetically (nominate subspecific taxa abbreviated as "nom."). Order Anseriformes (Waglet, 1831) Suborder Anseres Waglet, 1831 Family Anatidae Vigors, 1825 Subfamily Anatinae Swainson, 1837 Tribe Anatini Delacour and Mayr, 1945; Surface-feeding Ducks ß Subtribe Cairineae Boetticher, 1942; Long-billed Wood Ducks Supergenus Cairina Fleming, 1822; Greater Wood Ducks Genus Cairina Fleming, 1822; Muscovy Ducks C. moschata (Linnaeus, 1758); Muscovy Duck C. scutulata (Milller, 1842); White-winged Duck Genus Pteronetta Salvadori, 1895 P. hartlaubi (Cassin, 1859); Hartlaub's Duck Supergenus Aix Bole, 1828 Genus A/x Bole, 1828; Northern Wood Ducks A. sponsa (Linnaeus, 1758); American Wood Duck A. galericulata (Linnaeus, 1758); Mandarin Duck Subtribe Nettapodeae, new taxon; Stout-billed Wood Ducks Supergenus Chenonetta Brandt, 1836 Genus Chenonetta Brandt, 1836 C. jubata (Latham, 1801); Maned Duck Genus Nettapus Brandt, 1836; Pygmy-geese Subgenus Nettapus Brandt, 1836 N. auritus (Boddaert, 1783); African Pygmy-goose Subgenus Cheniscus Eyton, 1838; Pale-rumped Pygmy-geese N. coromandelianus (Gmelin, 1789); Cotton Pygmy-goose (nom., albipennis) N. pulchellus (Gould, 1842); Green Pygmy-goose Subtribe Anateae Boetticher, 1942; Dabbling Ducks Supergenus Amazonefta Boetticher, 1929; Micro-teal Genus Amazonefta Boetticher, 1929 A. brasiliensis (Gmelin, 1789); Brazilian Teal (nom., ipecutiri) Genus Callonetta Delacour, 1936 C. leucophrys (Vieillot, 1816); Ringed Teal Supergenus Lophonetta Riley, 1914; Proto-dabbling Ducks Genus Lophonetta Riley, 1914 L. specularioides (King, 1828); Crested Duck (nom., alticola) Genus Speculanas Boetticher, 1929 S. specularis (King, 1828); Bronze-winged Duck Supergenus Anas Linnaeus, 1758; True Dabbling Ducks Genus Mareca Stephens, 1824; Wigeons Subgenus Notonetta Roberts, 1922 M. capensis (Gmelin, 1789); Cape Teal Subgenus Chaulelasmus Bonaparte, 1838 M. strepera (Linnaeus, 1758); Gadwall (nom., couesi*) APPENDIX 2. Continued. Phylogeny of Dabbling Ducks 499 Subgenus Eunetta Bonaparte, 1856 M. falcata (Georgi, 1775); Falcated Duck Subgenus Mareca Stephens, 1824; True Wigeons M. sbilatrix (Poeppig, 1829); Chilo Wigeon M. penelope (Linnaeus, 1758); Eurasian Wigeon M. ameriana (Gmelin, 1789); American Wigeon Genus Anas Linnaeus, 1758; Typical Dabbling Ducks, sedis mutablis Subgenus Anas Linnaeus, 1758; Mallards Infragenus Melananas Roberts, 1922 A. sparsa Eyton, 1838; African Black Duck (nom., leu- costigma) Infragenus Anas Linnaeus, 1758; Northern Mallards, sedis mutabilis A. rubripes Brewster, 1902; American Black Duck A. fulvigula Ridgway, 1874; Mottled Duck (nom., macu 1osa) A. diazi Ridgeway, 1886; Mexican Duck A. platyrhynchos Linnaeus, 1758; Mallard (nom., conbos- chas) A. wyvilliana b Sclater, 1878; Hawaiian Duck A. laysanensis  Rothschild, 1892; Laysan Duck [A. oustaleti  Salvadori, 1894; Marianas Duck] infragenus Polionetta Oates, 1899; South Pacific Mallards A. luzonica Fraser, 1839; Philippine Duck A. superciliosa Gmelin, 1789; Pacific Gray Duck (nom., pelewensis, rogersi) A. poecilorhyncha Forster, 1781; Indonesian Spot-billed Duck (nom., baringtoni) A. zonorhyncha Swinhoe, 1845; Chinese Spot-billed Duck Infragenus Afranas Roberts, 1922; African Mallards A. undulata Dubois, 1839; Yellow-billed Duck (nom., ruppellO A. melleri Sclater, 1865; Meller's Duck Subgenus Spatula Boie, 1822; Blue-winged Ducks Infragenus Pterocyanea Bonaparte, 1841; Blue-winged Teal A. discors Linnaeus, 1766; Blue-winged Teal A. cyanoptera Vieillot, 1816; Cinnamon Teal (nom., or- inomus, borreroi, tropica, septentrionalium) Infragenus Spatula Boie, 1822; Shovelers A. smithii Hartert, 1891; Cape Shoveler A. platalea Vieillot, 1816; Red Shoveler A. rhynchotis Latham, 1801; Australasian Shoveler (nom., variegata) A. clypeata Linnaeus, 1758; Northern Shoveler Subgenus Nesonetta Gray, 1844; Australasian Teal Infragenus incertae sedis A. bernieri a Hartlaub, 1860; Madagascan Teal [infragenus Virago Newton, 1871; Gray Teal]  A. albogularis Hume, 1873; Andaman Teal A. gbberifrons Muller, 1842; Gray Teal (nom., gracilis, remissa  ) Infragenus Nesonetta Gray, 1844; Reddish Teal A. castanea Eyton, 1838; Chestnut Teal A. chlorotis (Gray, 1845); Brown Teal A. aucklandica (Gray, 1844); Flightless Teal (nom., nesi- otis) Subgenus Dafila Stephens, 1824; Pintails Infragenus Paecilonitta  Eyton, 1838; Pale-cheeked Pintails A. bahamensis Linnaeus, 1758; White-cheeked Pintail (nom., rubrirostris, galapagensis) A. erythrorhyncha Gemlin, 1789; Red-billed Pintail infragenus Dafilonettion Boetticher, 1937; Speckled Teal A. fiavirostris Vieillot, 1816; Yellow-billed Teal (nom., oxyptera) A. andlure (Sclater and Salvin, 1873);Andean Teal (nom., altipetans) Infragenus Dafila Stephens, 1824; Brown Pintails A. georgica Gmelin, 1789; Brown Pintail (nom., spinicau- da, niceforoi* ) A. acuta Linnaeus, 1758; Northern Pintail A. eatoni Sharpe, 1875; Eaton's Pintail (nom., drygalskii) Subgenus Querquedula Stephens, 1824; Holarctic Teal Infragenus Querquedula Stephens, 1824 A. querquedula Linnaeus, 1758; Garganey Teal Infragenus Nettion Kaup, 1829; Green-winged Teal A. formosa Georgi, 1775; Baikal Teal A. crecca Linnaeus, 1758; Common Green-winged Teal (nom., nimia) A. carolinensis Gmelin, 1789; American Green-winged Teal Subgenus Punanetta Bonaparte, 1856; Spotted Teal infragenus Punanetta Bonaparte, 1856; Pale-cheeked Teal A. versicolor (Vieillot, 1816); Silver Teal (nom., fretensis) A. puna (Tschudi, 1844); Puna Teal Infragenus Micronetta Roberts, 1922 A. hottentota (Eyton, 1838); Hottentot Teal Nine monotypic genera of waterfowl, included by some authors in the Anatini (or Cairinini), are here included in other tribes or subfamilies (Livezey 1986): Stictonetta naevosa to Stictonettinae; Plectropterus gambensis to Plectropterinae; Sarkidiornis melanotos, Malacorhynchus membranaceus, Hymenolaimus malacorhynchos, Merganetta armata, and Salvadorina waigiuensis to Tadorninae; and Marmaronetta angustirostris and Rhodonessa cary- ophyllacea to Aythyini. Assignable to lesser subgeneric taxon Horizonetta Oberholser, 1917, the North Pacific Mallards. Status tentative, possibly of hybrid origin (A. platyrhynchos x A. superciliosa; cf. Yamashina 1948); probably extinct. Position tentative. Possibly paraphyletic. Frequently used variants--Poecilonitta Gray 1840, Poecilonetta Reichenbach 1845, and Paecilonetta Bonaparte 1856--are junior synonyms. mmmmmmmmmmmmmmmmmmmmmmmmmmmmm mmmoommmmmmmmmmmmmmmmmmmmmmm