as well as amongst mammals, some variation, even amongst allied species, in the modification of these muscles; but this variation only occurs within comparatively narrow limits. In nine genera of Accipitres I have found that the inside or front plantar (flexor perforans digitorum) divides into three tendons at the foot of the tarsus, the three tendons leading to the 2nd, 3rd, and 4th digits respectively; whilst the outside or hind plantar (flexor longus hallucis) divides into two tendons at the foot of the tarsus, one leading to the hallux and the other to the 2nd digit. In the Osprey (Pandion haliaetus) this arrangement is modified; the front plantar is the same as in the nine other genera of Accipitres; but the hind plantar, instead of dividing into two, divides into four, and leads to each of the four digits. The Scansorial arrangement of the deep plantar tendons is very different from this; the front plantar leads to the 3rd digit only, whilst the hind plantar leads to the 1st, 2nd, and 4th digits. The heterodactyle arrangement differs, again, from all these modifications; the front plantar leads to the two front toes (the 3rd and 4th digits), whilst the hind plantar leads to the two hind toes (the 1st and 2nd digits). But all these apparently wide deviations from the common type may be generalized as follows:-The flexor perforans digitorum never leads to the hallux, and the flexor longus hallucis always does, except in those cases where the hallux is absent or so small that it does not receive any tendon from either of the deep plantars. In the Coraciiformes exactly the opposite is the case. hallux is always present, and always receives its tendon from the flexor perforans digitorum, and not from the flexor longus hallucis. The two plantars are always more or less coalesced, as they are in the Accipitres, Anseres, &c., but may easily be separated by gently tearing them asunder. In very many cases it is not necessary to do this, inasmuch as the tendon to the hallux branches off from the flexor perforans digitorum before the two plantars coalesce. The It is scarcely possible that such a remarkable condition of the deep plantar tendons could have been independently acquired by two groups of birds. It is now fifteen years since this unique arrangement of the deep plantar tendons in certain birds was recorded (Garrod, Proc. Zool. Soc. 1875, p. 344), but there are sufficient reasons to account for the small importance which was attached to the discovery. Two years before Garrod published his paper on the deep plantar tendons, he promulgated a new classification of birds, founded chiefly on the muscles of the thigh, the presence or absence of cæca, and the condition of the oil-gland, whether nude or tufted. Amongst the birds which possess the Coraciine arrangement of the deep plantar tendons are some with cæca and a nude oil-gland, some with no cæca and a tufted oil-gland, and others with no cæca and a nude oil-gland. There is also one group with an ambiens muscle, which none of the others possess. It is therefore quite obvious that Garrod could not give to this feature of the arrangement of the deep plantar tendons the importance which I have ventured to attach to it, without completely throwing overboard the main characters upon which his new classification was based. It does not appear that either Garrod or Forbes ever examined the deep plantar tendons of the Humming-birds. The Trochilidæ and the Cypselidæ constitute the Cypseliformes of Garrod, and the Macrochires of Nitzsch, and have been regarded by most ornithologists as closely allied families. Two important papers, however, have taken up a contrary opinion (Shufeldt, Proc. Zool. Soc. 1885, p. 886, and Journ. Linn. Soc., Zoology, 1889, p. 374), and in the latter the author mentions the arrangement of the deep plantar tendons. To verify the statement there made I have had the feet of three Humming-birds dissected by three independent experts, and in every case it has been found that the deep plantar tendons are not Coraciine, but very closely resemble those of the Eurylæmi. I propose therefore to remove the Trochili from the companionship of the Cypselidæ and place them as a suborder of the Pico-Passeres. If the Coraciiformes be regarded as one of the six subclasses into which the class Aves is divided (and the import ance of the character appears to warrant such a course), we may divide it into two orders, which may be diagnosed as follows: : MIMOGYPES.-No lateral bare tracts on the neck plantar leading to hallux; basipterygoid processes present; to which may be added, ambiens muscle present. This order contains the Cathartide of the New World. I have examined the feet of several species and find that the flexor perforans digitorum always leads to the 1st, 2nd, and 3rd digits, whilst the flexor longus hallucis sometimes leads to the 3rd and 4th, and sometimes to the 2nd, 3rd, and 4th digits, but never to the hallux. The Scansores having been disposed of amongst the PicoPasseres, and the Coccyges having been elevated to the rank of an order, the Picariæ, as thus restricted, appear to form a natural group, which may be diagnosed as follows : PICARIÆ. Hallux always present, and connected with the flexor perforans digitorum, and not with the flexor longus hallucis: no ambiens muscle. This order may be divided into three suborders, which may be diagnosed as follows: HALCYONES.-Front plantar leading to hallux; well-defined lateral bare tracts on the neck; spinal feather-tract not forked on the back to which may be added, palate desmognathous; vomer not ossified; basipterygoid processes absent. CORACIE. Front plantar leading to hallux; well-defined lateral bare tracts on the neck; spinal feather-tract forked on the upper back; oil-gland nude or absent. BUCEROTES. Front plantar leading to hallux; no lateral bare tracts on the neck; no basipterygoid processes; to which may be added, episternal processes perforated to receive the feet of the coracoids. It is possible that the Halcyones and the Coracia ought not to be divided upon their spinal feather-tracts, but upon the condition of their young at birth. The Caprimulgidæ are born covered with down. It is possible that this may also be the case with the Podargidæ and the Steatornithidæ. Leptosomus, with its powder-down patches like the Podargidæ, seems to have so many affinities with the Coraciidae that, in the absence of information as to the condition of their young at birth, I have not ventured to separate them. The Halcyones contain four families: Todida.-Halcyones with cæca and large tufts to the oilglands; two notches on each side of the posterior margin of the steruum. Momotida. Halcyones with no cæca; tufts of oil-gland very small or absent; one or both of the notches on each side of the posterior margin of the sternum converted into foramina. Coliida. Halcyones with neither central nor lateral bare tracts on the breast. Alcedinida. Halcyones with tufted oil-gland and no cæca; two notches on each side of the posterior margin of the sternum; well-defined central and lateral bare tracts on the breast. : The Coracia consist of seven families :- forwards. Caprimulgida.-Coracia with cæca and basipterygoid processes; dorsal vertebræ heterocœlous. Steatornithida.-Coracia with cæca and basipterygoid processes; dorsal vertebræ opisthocœlous. Podargide.-Coracia with a powder-down patch on each side of the rump; ten rectrices; no oil-gland. Leptosomida.-Coracia with a powder-down patch on each side of the rump; twelve rectrices; nude oil-gland. Coraciide.-Coracia with no powder-down patches; no basipterygoid processes; well-developed cæca; episternal processes not perfected to receive the feet of the coracoids. Meropida.-Coracia with the episternal process perforated to receive the feet of the coracoids. The Bucerotes consist of only one family, the Bucerotida. The importance of the character upon which I have ventured to found the subclass Coraciiformes has been to some extent admitted by some American ornithologists (cf. Stejneger, Stand. Nat. Hist. iv. p. 371), but it seems to me that it is not a natural arrangement to associate with the Coraciiformes birds with galline plantars, like the Cuculi, the Musophagi, and the Trochili; birds with Passerine plantars, like the Upupa; birds with Picine plantars, like the Scansores; and birds with Trogonine plantars, like the Trogones; whilst birds with Coraciine plantars, like the Cathartes, are excluded. This is, of course, a matter of opinion; but to include the Meropida and the Bucerotida in the Halcyones (or Alcedinoideæ, op. cit. p. 395), which are described as having the "dorsal tract simple between the shoulders," appears to me to be inconsistent with fact. So far as I can judge from an examination of specimens of three species of Merops, the spinal feather-tract is forked between the shoulders in the Meropidæ, though perhaps not quite so far forward as in the Coraciidæ, whilst the spinal feather-tract of the Bucerotide is described by Nitzsch as double in some species and simple in others on the interscapular region. After the Bucerotida and the Meropida have been removed from the Halcyones, the remaining families can be described as not having the spinal feather-tract forked either on the upper or lower back, a character which greatly simplifies their diagnosis. In conclusion, I take the opportunity of correcting an unfortunate blunder in my paper on the Pico-Passeres (Ibis, 1890, p. 36). At the bottom of the page the names Bucconidæ and Galbulidæ should be transposed. XIX.-Descriptions of three new Species of Flycatchers. By R. BOWDLER SHARPE, F.L.S., F.Z.S., &c. AMONGST a few birds presented to the British Museum by Dr. G. Vorderman is an example of a small Siphia from Mount Gedeh, in Java, which appears to me to be undescribed. It belongs to the plain-coloured group of the genus, and its nearest ally seems to be the female of Siphia cæruleata (Cat. B. iv. p. 457), which has also whitish legs, but which differs in having a blue tail. |