ORDER COLUMB^E. 241 



a^ in the Galliformes the postfrontal process fuses with the zygoma. The cervical 

 vertebrae are fifteen to seventeen in number, the sternum having its posterior margin 

 two notched on each side, or a single notch and a foramen on each side, this resulting 

 in different groups. 



The carotids are variable, but the tracheo-bronchial syrinx shows little variation, 

 the intrinsic muscles always present, the extrinsic asymmetrical, while the digestive 

 system is pericoelous, showing gradation from the typically pericoelous to the typically 

 mesogyrous, caeca small and nipple-like or absent. The leg muscles are variable, 

 the complete formula ABXY+ being common, sometimes the ambiens is missing, 

 and in the genus Lopholaimus the accessory femoro-caudal absent. The biceps slip 

 is present of peculiar formation, and the expansor secundariorum sometimes 

 absent though generally present. The oil gland is sometimes present, sometimes 

 absent, but when present never feathered ; this variation is seen in closely allied 

 species. The aftershaft is always said to be absent but the wing, though generally 

 aquincubital has been recorded in one instance as quincubital. The pterylosis, as. 

 to detail, has not yet been reported upon. In the Raphiformes the scapula and 

 coracoid are anchylosed, but in the Columbiformes the coracoids are free and though 

 they meet they do not overlap ; in both the furcula has no hypocleidium. It is 

 probable that the association of the extinct flightless forms, Pezophaps and Eaphus 

 ( = Didus olim) is incorrect, palaeontologists having been misled by items of con- 

 vergence, not of phylogenetic import. 



In this place may be mentioned the Sand Grouse, which we have allowed as 

 of ordinal rank between the Galli and the Columbae. Externally they agree better 

 to our eyes with the former, and we cannot understand those who have considered 

 them as Pigeon -like. Internally they show many characters of Galline aspect and 

 a few approaching those of the Pigeons, and hence have been regarded as a living 

 representative of the Galline ancestral form from which the Pigeons developed. We 

 cannot see, however, that any other than ordinal rank must be accorded the three 

 groups. 



SUBORDER COLUMBIFORMES. 



The superficial homogeneity of this group has made the acceptance of families 

 and even genera a difficult task, and not the least assistance, but rather the reverse, 

 has resulted from the criticism of the so-called deeper-seated characters. As a matter 

 of fact, we find comparatively good superficial features but comparatively valueless 

 deeper-seated ones in connection with the groups. 



The internal features have been synopsised above, but we may here add that 

 Garrod's differentiation was based on the presence or absence of the ambiens, caeca, 

 oil gland and gall bladder, and by usage of this set of items he classed Carpophaga 

 with Phaps and then Phlogoenas with Ptilopus in different families. It may be 

 that Carpophaga and Ptilopus have descended from different sources, and that 

 Phaps and Phlogcenas (of which Gallicolumba seems to be the correct name) are 

 also homeomorphic, but it would require more evidence than the above set of items 

 to prove it. As a matter of fact, the most aberrant form, judged from the superficies, 

 proves to be the most peculiar when internal features are examined. We have 

 allowed six families, Columbidae, Turturidae, Treronidae, Microgouridae, Gouridae, 

 and Didunculidae. The first named includes the Palaearctic forms, and one or two 

 of the second family also range into the Palaearctic but the headquarters are tropical ; 

 the remaining four families are Oriental and Papuasian -Australian in range, the last 

 three being represented by single species, but not Australian. Three families are 

 therefore regarded as being represented in Australia the Columbidae, Turturidae 

 and Treronidae and we are of the opinion that this arrangement is imperfect, but 

 we cannot at this time prove it. In America the first two families occur, and 



R 



