'# " %( + &" & ,

Procellariiformes. Pelecanoididae (diving-petrels) are probably nested within Procellariidae (petrels and shearwaters) (Cracraft et al. 2004), so I show them as sister taxa in Fig. 1.2. This is the only resolved node in my reanalysis (Harshman 1994) of Sibley and Ahlquist’s (1990) data. Hydrobatidae may be polyphyletic (see below) and this prevents any further resolution among families. Pelecaniformes. This order is divided into three pieces. Phaethontidae (tropicbirds) is apparently not closely related to any of the other families. The relationships of Pelecanidae (pelicans) have been noted above. This leaves a clade I will call “core pelecaniforms” despite the absence of pelicans from the group, composed of Fregatidae (frigatebirds), Sulidae (boobies), Phalacrocoracidae (cormorants), and Anhingidae (darters). The only resolved node in my reanalysis (Harshman 1994) of Sibley and Ahlquist’s (1990) data unites these four families. Siegel-Causey (1997) and Cracraft et al. (2004, fig. 27.4) also find this grouping. In other studies (van Tuinen et al. 2001; Cracraft et al. 2004, fig. 27.7) only the last three are united or Fregatidae is only weakly attached to the others (Cracraft et al. 2004, fig. 27.6). Relationships among Sulidae, Phalacrocoracidae, and Anhingidae are ambiguous, with either phalacrocoracids (van Tuinen et al. 2001; Fain and Houde 2004) or sulids (Siegel-Causey 1997; Cracraft et al. 2004) sister to the other two. Falconiformes. Though the order cannot be confidently asserted as monophyletic, relationships are well established among families of the suborder Accipitres, with Sagittariidae (secretary bird) the sister of Pandionidae (osprey) and Accipitridae (hawks) (Cracraft et al. 2004; Fain and Houde 2004); but see Griffiths (1994). Gruiformes. The relationships of Otididae (bustards), Cariamidae (seriemas), and Mesitornithidae (mesites) to any other families (gruiform or otherwise) cannot be determined at present. Turnicidae (buttonquail) belongs to Charadriiformes (see above). Eurypygidae (sunbittern) and Rhynochetidae (kagu) are sister taxa (Houde et al. 1997; Livezey 1998; Cracraft et al. 2004; Fain and Houde 2004), but their relationships to other families cannot be determined. This leaves a “core gruiforms” consisting of Gruidae (cranes), Aramidae (limpkin), Psophiidae (trumpeters), Rallidae (rails), and Heliornithidae (finfoots), for which there is strong support (Houde et al. 1997; Livezey 1998; Cracraft et al. 2004; Fain and Houde 2004). There is strong support for a sister group relationship between Gruidae and Aramidae and between Rallidae and Heliornithidae (Houde 1994; Houde et al. 1997; Livezey 1998; Cracraft et al. 2004; Fain and Houde 2004). The placement of Psophiidae is disputed, but the only strong support is for a position as sister of Gruidae plus Aramidae (Livezey 1998; Fain and Houde 2004). Charadriiformes. All molecular analyses (Sibley and Ahlquist 1990; Ericson et al. 2003; Paton et al. 2003; Fain and Houde 2004) are agreed on the

relationships of families within Charadriiformes. Paton et al. (2003) discuss incongruence between their tree and that of Sibley and Ahlquist (1990), but they are referring there to the Tapestry, while I refer here to the more rigorous Fitch tree (Sibley and Ahlquist 1990; Harshman 1994). The order is traditionally divided into three suborders, Charadrii, Lari, and Alcae (del Hoyo et al. 1996), but Alcae is nested within Lari, and the traditional Charadrii is polyphyletic. The three basal clades do consist of plover-like birds, gull-like birds, and sandpiper-like birds, which might be given the names Charadrii, Lari, and Scolopaci, respectively, with the first being sister to the other two. Ibidorhynchidae (ibisbill) has never been sampled and cannot be placed. Lacking any better notion, I have placed it at the basal node of Charadriiformes. Caprimulgiformes. There is no study strongly supporting the relationship of any caprimulgiform family to any other (Harshman 1994; Mariaux and Braun 1996; Fidler et al. 2004). The only supported relationship is that of Aegothelidae to Apodiformes, as discussed above. Apodiformes. Remarkably few studies have included representatives of all three families of apodiforms, but those that did have unsurprisingly found Apodidae (swifts) and Hemiprocnidae (treeswifts) to be sister taxa (Johansson et al. 2001; Mayr 2002a; Mayr et al. 2003; Chubb 2004a). Coraciiformes. Relationships within “core coraciiforms” are partially resolved. One clade is composed of Alcedinidae, Todidae, and Momotidae, though relationships among these families are contradictory (Sibley and Ahlquist 1990; Harshman 1994; Espinosa de los Monteros 2000; Johansson et al. 2001; Johansson and Ericson 2003), and another clade of Coraciidae and Brachypteracidae (Johansson et al. 2001; Kirchman et al. 2001; Johansson and Ericson 2003). The position of Meropidae is unresolved, though there is some evidence that it is sister to the first clade (Johansson and Ericson 2003). There is another clade of traditional coraciiforms whose relationships to core coraciiforms are unresolved; it consists of Bucerotidae, Upupidae, and Phoeniculidae, in which the last two are sister taxa (Sibley and Ahlquist 1990; Harshman 1994; Johansson et al. 2001; Cracraft et al. 2004). Leptosomidae certainly does not belong to the roller clade and probably does not belong to the core coraciiforms (Kirchman et al. 2001). Piciformes. There is a basal split between Capitonidae (which includes Ramphastidae) and the other two families, Indicatoridae and Picidae (Sibley and Ahlquist 1990; Harshman 1994; Johansson and Ericson 2003; Cracraft et al. 2004).