Feathers have been found at Clutha River, near Roxburgh, and also in caves near Queenstown. Those from Clutha are mostly dark, being black with white tips; while the Queenstown ones resemble feathers of Apteryx australis in colours. Professor Owen has shown that Megalapteryx huttoni was feathered down to the toes, and in the plate I have represented it clothed with feathers similar to the Clutha ones, which I believe belong to this species. The Moas at one time must have been extraordinarily numerous, both in numbers and species, and they varied in height from 2½ feet to 12 feet. Professor Parker has shown that some of the species had crests of long feathers on the head, and, as some adult skulls of the same forms show no signs of this, he infers that the males alone had this appendage. There has been much discussion as to the time when the Moas became extinct, and we know for certain that the two species, Dinornis maximus and Anomalopteryx antiquus, belong to a much earlier geological epoch than the bulk of the other species. It would be too lengthy for my purpose to go into the arguments, but we can, by the study of the "kitchen middens" of Maoris and their traditions, fairly adduce that the Maoris arrived in the North Island some 600 years ago, that they hunted Moas, and that they exterminated them about 100 to 150 years after their arrival. In the South, or rather Central, Island, the Maoris appear to have arrived about 100 years later, and to have exterminated the Moas about 350 years ago. It is only fair to say, however, that Monsieur de Quatrefages adduces evidence in his paper which goes far to prove that Moas existed down to the end of the 18th or even beginning of the 19th century in those parts of the Middle Island not, or scantily, inhabited by Maoris.
The Dinornithidae form a separate group of the order Ratitae, in no way closely related to the Australian Emu (Dromaius), as many ornithologists have asserted, but nearer to the South American Nandu (Rhea) than any other living Ratitae, though exhibiting many characters in common with the Apterygidae. There have been a number of classifications set up of this family. The first by Reichenbach, in 1850, with 7 species and 7 genera!
The next was by Von Haast, in 1873, who enumerated 10 species, divided into 4 genera. The third was Lydekker's, in 1891, who acknowledged 23 species, divided into 5 genera. Then came Hutton's, in 1892, which left out Megalapteryx, with its then known 2 species, and acknowledged 26 species, divided into 7 genera. Lastly we have Professor Parker's, in 1895, in which again Megalapteryx is left out, and 21 species are acknowledged, divided into 5 genera. There has been a great amount of controversy as to the number of species of Moas which really ought to be distinguished, and of late years there has been a tendency to unite most of the species as synonyms, the authors declaring that bones vary to such a degree that all the characters relied on for the distinguishing of the various species were individual variations, and that, besides, it was impossible that so many distinct forms could have occurred in such a small area. The extreme of this lumping was reached when Professor Forbes, in the Bulletin of the Liverpool Museums, III, pp. 27 and 28 (1900), divided the Moas into six genera, each with a single species. He thus ignores the fact that by doing so he has united forms which were founded on FULLY ADULT bones, and yet some of them were only about half or two-thirds the size of the others. I personally think that too many species have been made, and at least 7 of Captain Hutton's forms must be sunk. On the other hand some have been described since 1895 and 1900, and I have been obliged to name others rather against my will, so that in spite of uniting so many species of others I find I am obliged to acknowledge more species than anyone else. I have divided these into genera according to Professor Parker's classification, only adding Palaeocasuarius of Forbes, with 3 species, and Megalapteryx, with 5, which brings my number up to 38 species, divided into 7 genera. My reasons for not uniting these into 7 species and 7 genera, as those of the "lumping school" do, are twofold,—first, the bones of the Ratitae are much more solid than those of other birds, and are not given to so much individual variation; and, secondly, in the face of the great number of species of Paradise Birds and Cassowaries found on New Guinea, the contention that there could not be so many species of Moa on so small an area is not easily maintained. Moreover, we have strong support in the present fauna and flora for the presumption that, when the Moas first came into existence and differentiated into species, New Zealand was a much larger area, stretching at least from the Macquarie Islands in the south to the Kermadecs in the north, and from Lord Howe's Island on the west to the Chatham Islands on the east. So that, like the giant tortoises on the Galápagos Islands,
they only got driven so closely together after their specific differentiation, when the land gradually subsided, owing to volcanic action. The differentiation of the family is as follows:—
DINORNITHIDAE.
Skull with a short and wide beak. Pectoral girdle very small or absent, wing absent, only an indication in Dinornis dromioides. Hallux absent or present. An extension bridge to the tibio-tarsus, which is placed near the inner border of the bone. No superior notch to the sternum. Most of the species of very large size. The tarso-metatarsus is either long and slender or short and wide, and its anterior surface may or may not be grooved. The second trochlea is longer than the fourth, the third is not pedunculated, and there is no perforation in the groove between the third and fourth trochlea. In the tibio-tarsus the cnemial crest rises well above the head; the extensor groove is separated by a considerable interval from the inner border of the bone. There is a well-defined intercondylar tubercle; the intercondylar gorge is deep, and there is no deep pit on the lateral surface of the entocondyle. The femur may be either slender or stout, but is not markedly curved forwards. The popliteal depression is deep, and the summit of the great trochanter rises considerably above the level of the head. The pelvis approximates to that of the Apterygidae, but the pectineal process of the pubis is less developed, and the ischium and pubis may be longer and more slender. The coracoid and scapula are aborted and may be absent. The sternum, which may be either long and narrow, or broad and short, differs from that of the Apterygidae by the absence of the superior notch, the divergent lateral processes, and the reduction of the coracoidal grooves to small facets or their total disappearance. The cervical vertebrae are relatively short, an expanded neural platform as far as the sixth.
In Anomalopteryx and Megalapteryx the number of cervical vertebrae is 21, and there are 2 cervico-dorsal and 4 free dorsal vertebrae, so it is fair to assume that this is the correct number throughout the family.
The feathers had after-shafts.
I have adopted Professor Parker's classification in the genera, only substituting Cela Reichenbach for Mesapteryx Hutton, which is a synonym of Megalapteryx Haast. As to the species I have used my own judgment; I felt obliged to name a number of species acknowledged by Parker and Lydekker but not named, because this system of indicating species by the letters A, B, C, &c., which has crept into our nomenclature, will make all understanding impossible, as not always the same species is denoted by the same letter. A few of these species will naturally later have to be sunk, as some have been founded on skulls and others on leg bones, or so, which, when we get perfect individual skeletons may prove to be identical, but I do not think these will be many.