has mentioned the existence among the Striges of a 'scapula accessoria.' The coracoids are not in contact at their articulation with sternum; the procoracoids are moderately large, and the clavicle reaches both them and the scapula.

If it were not for Photodilus, it might be possible to divide the Striges into two families, Strigidæ and Bubonidæ. As it is, it may perhaps be permissible to regard the order as containing but one family, but two sub-families, viz. Striginæ

好

FIG. 129. SYRINX OF Scops leucotis (AFTER BEDDARD).

FIG. 130.-SYRINX OF Bubo (AFTER BEDDARD).

and Buboninæ, to which possibly a third, Photodilinæ, might be added.

The syrinx of the Striges has been chiefly described by WUNDERLICH' and by myself. This group is one of the few that present the remarkable variety of the voice organ which has been termed the bronchial syrinx. All the owls, so far as they have been examined, possess one pair of intrinsic muscles and the usual one pair of extrinsic muscles. Scops leucotis has the most modified syrinx. In this bird (see fig. 129) the intrinsic muscles are attached so far down the bronchus as to the tenth bronchial ring, and, as will be

''Beiträge zur vergleichenden Anatomie und Entwicklungsgeschichte des unteren Kehlkopfes der Vögel,' Nov. Act. Leop. Akad. xlviii. 1884, p. 1. 2. On the Classification of the Striges,' Ibis (5), vi. p. 355.

24

seen from the figure, the bronchial rings in front of this attachment are complete rings, with no membranous interspace left. In Strix, on the other hand, and in Bubo and Syrnium, the intrinsic muscles are inserted on to the first bronchial semi-ring. In Asio seven complete rings intervene between the bifurcation of the trachea and the first incomplete bronchial semi-ring, to which the muscles are attached. In Photodilus the intrinsic muscles are inserted on to the second bronchial semi-ring.

Until lately the owls have been almost invariably placed in the immediate neighbourhood of the diurnal birds of prey. Latterly, however, the opinion has been gaining ground that it is to the picarian birds (in a wide sense) that they are most nearly allied. This opinion, more than hinted at by GARROD and NEWTON, has been given a practical shape in the classifications of FÜRBRINGER and GADOW. The latter has ingeniously pointed out that it is impossible to imagine that the Striges have been derived from the Accipitres, since, although without an ambiens, they have much the same structure of foot as the Accipitres with an ambiens. Hence it is difficult to believe that they would have lost it; he concludes that they are derived from some bird without an ambiens, and the failure of MITCHELL to find the last trace of the missing ambiens-obvious in some birds which are clearly the descendants of birds with an ambiens-still further supports that way of looking at the matter. Even in the skull, where the principal likenesses between the Accipitres diurnæ and nocturna (as the two groups in question have been called) have been seen, there are really many differences. It is only, for example, in the skulls of those Accipitres to which the owls have been supposed to have the least resemblance, i.e. the Cathartidae and Serpentariidæ, that there are basipterygoid processes. The owls are decidedly not desmognathous (in the sense of a maxillopalatine union), and their lacrymal is quite different from that of the hawks and eagles. The palate, too, is incomplete in front of the maxillo-palatines, not solid, as in the Accipitres. As to other anatomical features, it is harder to

find likenesses than differences. The Accipitres have rudimentary cæca, a biceps slip, the expansor secundariorum, a tufted oil gland, an aftershaft (except Pandion); the deep flexor tendons are different, and, in short, the differences are as great as those which separate any two groups of carinate birds.

PSITTACI

Definition.-Twelve rectrices; 2 aftershaft present; aquintocubital; zygodactyle. Skull desmognathous, holorhinal, without basipterygoid processes. Biceps slip and expansor secundariorum absent.3 Muscle formula, AXY+ or -. Νο cæca; a crop present.

The parrots are an almost cosmopolitan group, being most abundant, however, in the tropics. Count SALVADORI, in his British Museum catalogue of the group, allows five hundred species, distributed among seventy-nine genera. The parrots are a very sharply defined group, there being no dubious outlying forms. They are usually brilliantly coloured, and lay white eggs in hollows of trees. With the exception of the owl parrakeet (Stringops) of New Zealand the parrots are arboreal birds, as, indeed, the zygodactyle feet denote. As to external characters, the exaggeratedly hawklike bill is well known. The almost universal twelve rectrices distinguish the group, but in other external and internal characters the parrots show considerable diversity of structure, as is sometimes the case with large and widely distributed groups; compare, for instance, the pigeons, which present many other analogies to the parrots.

The oil gland is a structure which may be wanting or developed. The table on p. 268 indicates some of the genera

See also under Caprimulgi,' p. 243.

With the sole exception (cf. GADOW) of Oreopsittacus Arfaki.

3 See below, p. 261.

External characters and many other points in the anatomy of parrots are dealt with by GARROD, On some Points in the Anatomy of the Parrots,' &c., P. Z. S. 1874, p. 247, and Notes on the Anatomy of certain Parrots,' ibid. 1876, p. 691; see also FORBES, On the Systematic Position of the Genus Lathamus,' ibid. 1879, p. 166.

in which it is present or absent. When present it is invariably tufted, and generally of fair size. In Cacatua sulphurea, however, the oil gland, though present, is small, and has but a single small down feather upon it.

The parrots are a group of birds which agree with the Accipitres in that some genera have powder-down patches while others have not. The table already referred to indicates the facts so far as they have been ascertained. The degree of development of the powder-downs, however, differs considerably, though in no parrot is there more than a single pair of definite powder-downs which are lumbar in position. In Cacatua sulphurea, for instance, there are a pair of such patches, one on either side of the dorsal tract. These send up a few scattered powder-down feathers as far forward as the neck, and a few to carinal spaces and between the branches of the ventral tract.

In Calopsitta Nova Hollandia there are the same lumbar patches of a reniform outline; but the powder-downs are entirely confined to this region of the body.

Calyptorhynchus stellatus is more like Cacatua, but the lumbar patch is not so well developed.

In Psittacula passerina there are lumbar patches more elongated but narrower than those of the parrots already referred to; there are also scattered powder-downs not aggregated into patches.

Brotogerys tirica has no defined patches, but simply a few scattered powder-downs, which, however, are more numerous in the lumbar regions. In Coracopsis, Chrysotis, and Pionus there are the same generally diffused powder-down feathers not aggregated into definite tracts. The same may be said generally of Psittacus, though such powder-downs as there are are limited to the lumbar region.

The general pterylosis of the parrots is as follows: From the general covering of the head arises a narrow dorsal tract, which bifurcates in the interscapular region. Between the arms of this fork are the arms of another fork, which unite near the oil gland to form the single straight, short posterior part of the dorsal tract. On the ventral surface the tract is

single, or double on the neck, and where it widens out on either side of the carina sterni a strongly feathered lateral branch is given off.

The variations shown in the pterylosis are not great. They concern the more or less definite separation of the lateral ventral tract and the slighter or more pronounced feathering of the anterior end of the posterior dorsal Y. Thus in Lathamus and Platycercus the lateral ventral tracts are well marked, and the posterior fork of the dorsal tract does not diminish much in width where it comes into contact with the anterior fork. In Trichoglossus the exact reverse of these conditions obtains, the lateral ventral tracts being but obscurely delimited from the main tract, and the dorsal tracts of the hinder part of the body almost ceasing before their junction with the anterior half.

Psephotus, Cyanorhamphus, Pyrrhulopsis, Agapornis, &c., agree with Platycercus. In Ara the outer branch of the pectoral tract is not definitely separable, but the dorsal tracts are more like those of Platycercus. Conurus is much the

same.

The syrinx of the Psittaci1 seems to show two main varieties.

a. In the following species there is a syrinx of the type which will be described immediately:

:

The syrinx is in these species remarkable for the fact that the first semi-rings of the bronchi are weak and cartilaginous, and are usually separated from each other by considerable tracts of membrane. Cacatua itself represents the most ex

The syrinx has been chiefly studied by GIEBEL, Zur Anatomie der Papageien,' Zeitschr. f. d. ges. Wiss. xix. p. 133, and by PARSONS and myself, * On certain Points in the Anatomy of Parrots,' &c., P. Z. S. 1893, p. 507.