xviii INTRODUCTION. 



not enclose any ephippial space, therein differing from the Crow's, and resembling the typical Starling's. 

 There are nineteen remiges, of which ten are on the hand ; they increase in size up to the fifth. The rectrices 



are twelve in number. The oil-gland is nude The gizzard is well developed. The intestines are 1G inches 



long, with the bile-ducts 2| inches from the gizzard. The caeca are 1 inch from the cloaca and \ inch long, being 



cylindrical. There is one carotid artery, the left The palate is strictly segithognathoxis ; that is, the 



vomer is truncate in front abruptly, and cleft behind ; the postero-external angles of the palatines are produced ; 

 the maxillo-palatines are slender, and approach towards, but do not unite with, one another, nor with the 

 vomer, which they partly embrace. There is no ossification in the nasal septum anterior to the vomer. The 

 whole cranial configuration closely resembles that of Sturnus; but the mandible, instead of being bent upwards, 

 is straight. Like it, the palatines are narrow and approximate ; the antcro-internal angles of the posterior 

 portions of those bones are reduced and rounded off, as is sometimes the case with Sturnus. The vomer is 

 completely truncated in front, and is not prolonged forwards at its external angles, as in Corvus and its allies. 

 The zygoma is not so slender as in Sturnus ; but the curves are similar. The articular surfaces on the 

 quadrate bone for the mandible are proportionally very large. The anterior extremities of the pterygoid bones 

 articulate with the sphenoidal rostrum much as in Corvus, meeting in the middle line behind the posterior 

 extremities of the palatines for a short distance. The maxillo-palatines, in their approximate portions, are 

 shorter from before backwards than in Sturnus, and much resemble those of Corvus. The anteroinferior 

 processes of the orbit are large and spongy; they almost touch the zygoma. But the most characteristic 

 portion of the skull of Heteralocha is the occipital region ; and in this it presents a great exaggeration of the 

 peculiarities of Sturnus and its allies. In Corvus and most Passerines the digastric muscles occupy a narrow 

 space intervening between the auditory meatus and the mass of occipital muscles, not extending so high up 

 the skull as the latter. The occipital ridge encloses a space elongated from side to side, and of but little depth. 

 In Sturnus the digastrics are much broader, and they narrow the occipital space ; they also extend up the 

 skull to so great an extent that they nearly meet in the middle line above the origin of the biventres cervicis 

 muscles ; but in Heteralocha they are of still greater size, and, meeting above the middle line, they form a 

 strong ridge, which extends for some distance into the parietal region vertically. This peculiar development 

 of these muscles produces a corresponding change in the shape of the space enclosed by the occipital ridge. 

 In Heteralocha it is almost circular, and it extends some way above the foramen magnum. In Sturnus there 

 is an approximation to this condition. A vertical parieto-occipital ridge in many other birds closely resembles 

 that of Heteralocha ; but it is the median limit of the temporal fossa in most. Correlated with this extensive 

 digastric origin is a large surface for its insertion. The angle of the mandible is prolonged directly back- 

 wards for this purpose, in a manner unique among Passerine birds, but well seen in the Anatidae. In Sturnus 



the angle of the mandible is slightly prolonged backwards for a simdar purpose In the sternum 



Heteralocha differs in no important point from Sturnus, except that the posterior notches tend to be converted 



into foramina, as observed by Mr. Eyton in his 'Osteologia Avium/ In conclusion, it may be stated 



that the anatomy of Heteralocha shows clearly that it is truly Passerine, and not related to Upupa, as was 

 previously supposed by most authors. When examined more in detail its relation to the Sturnidm is found to 

 be very intimate, and its structure is clearly not closely allied to that of the Corvidce. In its relation to 

 Sturnus it seems to present an exaggeration of the peculiarities of that bird, which would place it at the head 

 of the family"*. 



Fam. B.allid.£. In a country possessing such forms as Notornis and Porphyrio we might 

 naturally look for the occurrence also of Tribonyx. Both of the latter are known to have a wide 

 geographic range, while Notomis, which is a strictly local form, appears to combine in some 

 measure the characters of each, being allied to Porphyrio in the form of its bill and in its 



* P.Z.S. 1872, Part ii. pp. 643-647. 



