FIG. 201.-CONVOLUTED WINDPIPE OF Tantalus ibis (AFTER GARROD). c, coracoid; f, furcula; st, sternum; r.b, l.b, bronchi.

are a little more complicated. The tendon of the brevis is. obscurely divided into two from nearly its commencement ; the more anterior of these again divides into two, one of which runs forward to be inserted on to the fore arm separately from the hinder part, which remains continuous with the rest of the tendon; there is a recurrent slip to longus.

In Abdimia the tendons are much the same; in both these genera the propatagialis pectoralis of FÜRBRINGER is muscular.

In Tantalus leucocephalus the broad fascia-like tendon of the brevis gives off a wristward slip, from the junction of which with tendon of extensor matacarpi radialis a patagial fan arises.

The pterylosis of the Ciconiida has been studied by NITZSCH. The neck is continuously feathered down to about the middle, where the spinal and ventral tracts respectively become divided into two. The two spinal tracts are narrow but strongly feathered, and cease abruptly at about the end of the scapula; after a short space they recommence as a bifid but feebly feathered tract, the limbs of which unite a little way in front of the oil gland. The ventral tracts are broad upon the pectoral region but narrow towards the vent.

In Pseudotantalus leucocephalus NITZSCH could not find the aftershaft, but nevertheless one appears to exist.

In Mycteria and Leptoptilus argala the dorsal tract has posteriorly no spinal space.

In L. argala NITZSCH states the aftershaft to be absent. The oil gland has two apertures on each side in Ciconia alba, five in C. nigra; L. argala has no less than six. Anastomus coromandelicus has three; Tantalus leucocephalus has the same number.

The deep flexor tendons of the Ciconiida are constructed on the plan of type I. In Tantalus leucocephalus there is a slight variation; a small vinculum runs to flexor communis before the latter divides into three, and then a broader vinculum, chiefly going to tendon of digit II., but also slightly to III. and IV., binds together the two tendons. In C.

nigra there is the ordinary vinculum and a special slip to digit II.

The lungs in the Ciconiidæ are at least often distinguishable from those of the Ardeidae by the great deficiency of the muscles arising from the ribs and attached to the pulmonary aponeurosis. In Cancroma there were four pairs of such muscles, arising from a corresponding number of ribs, in Nycticorax five; but in Ciconia alba I only found one pair inserted on to the aponeurosis in front of the septum bounding the anterior intermediate air sac anteriorly. Professor WELDON found no such muscles in a considerable number of storks.

The tongue is always small. The proventriculus is zonary. The liver is nearly equilobed, and there is always a gall bladder. The intestinal measurements of a number of species are as follows:

The number of cervical vertebræ is seventeen (Leptoptilus, Tantalus) or eighteen (Xenorhynchus); the hypapophyses are feeble; there is a ventral canal formed by union of catapophyses of C7-C11 (Xenorhynchus), C7-C12 (Tantalus). Four ribs reach the sternum in Xenorhynchus, five in the others. The sternum has one pair of notches, and the spina externa is absent or small. The procoracoid is of fair size, but does not reach the clavicle; the coracoids meet

The hypoclei

at their sternal insertion in Leptoptilus.' dium articulates with carina sterni. The skull is desmognathous, holorhinal, and without basipterygoid processes; there are in Tantalus rudiments of these processes in the shape of a minute spine. The holorhinal character of the nostrils is largely marked by ossifications of the alinasals; the nostrils are thus much reduced in size, in a fashion suggestive of the Steganopodes and possibly significant, but there is no bony septum between them. This distinguishes the storks from the herons, as does also the form of the palatines. These bones are in the first place not cut off at right angles behind, as in the herons, while the internal lamina only bounds the interparietal space, and is at most (Xenorhynchus) carried back to the end of the bones as a slight median keel. This is absent in other storks. In Xenorhynchus and Leptoptilus the palatines again approach each other, and are only separated by the vomers just behind the maxillo-palatines. Opposite to this point each palatine is produced into a strong outwardly directed snag, large in Xenorhynchus, hardly indicated in other storks. The interorbital septum is entire. The large lacrymal is perforated or deeply notched for duct of gland.

Family Ardeida.-The herons contrast with the storks in (1) the tracheo-bronchial syrinx always furnished with a pair of intrinsic muscles, (2) non-division of pectoralis primus into two layers, (3) invariable absence of ambiens, (4) presence of powder-down patches, (5) absence or weakness of vinculum, (6) presence of a single cæcum only.

On the other hand the two families agree in (1) absence of biceps slip to patagium, (2) presence of expansor secundariorum, in addition, of course, to the points enumerated in the definition of the group.

In Cancroma the oil gland is nude. The herons have four or six powder-down tracts. Six are found in Cancroma, Butorides atricapillus, Ardea cocoi, and other species;

Not in Xenorhynchus and Dissura; there is a trace of an overlap in Abdimia, Tantalus, and Platalea.

four in Ardetta and Botaurus. The number of rectrices differs much; there are only eight in Ardetta exilis and A. involucris, ten in Botaurus (not invariably), twelve in Ardea cocoi and Cancroma cochlearia.

The pterylosis differs from that of the storks by the narrowness of the tracts. The spinal space begins with the commencement of the neck, and only terminates a little way in front of the tufted oil gland. The anterior part of the spinal tract is not always more strongly feathered than the posterior part, and there is (according to the figures of NITZSCH) hardly a break between them. The ventral tracts also are separate early on the neck. In Cancroma they divide on the breast into a broader, stronger outer tract, which ceases just below the metapatagium, and a narrow inner tract. The anterior ventral powder-down patches constantly interrupt the continuity of the anterior and posterior sections of the ventral tracts.

A curious absence of any apparent relationship between the relative length of the sections of the alimentary canal, and of the alimentary canal as a whole, and of the food, is shown by the above table, drawn up by Mr. GARROD, which I reproduce here from his notes.

They seem, however, to ally the last four genera, which