English as their mother-tongue. Still, under ordinary circumstances, connexion of speech does indicate more or less connexion of ancestral race. As a guide in tracing the history of civilization, language gives still better evidence, for common language to a great extent involves common culture. The race dominant enough to maintain or impose its language, usually more or less maintains or imposes its civilization also. Thus the common descent of the languages of Hindus, Greeks, and Teutons is no doubt due in great measure to common ancestry, but is still more closely bound up with a common social and intellectual history, with what Professor Max Müller well calls their 'spiritual relationship.' The wonderful permanence of language often enables us to detect among remotely ancient and distant tribes the traces of connected civilization. How, on such grounds, do savage and civilized tribes appear to stand related, within the various groups of mankind. connected historically by the possession of kindred languages?

The Semitic family, which represents one of the oldest known civilizations of the world, includes Arabs, Jews, Phoenicians, Syrians, &c., and has an earlier as well as a later connexion in North Africa. This family takes in some rude tribes, but none which would be classed as savages. The Aryan family has existed in Asia and Europe certainly for many thousand years, and there are well-known and well-marked traces of its early barbaric condition, which has perhaps survived with least change among secluded tribes in the valleys of the Hindu Kush and Himalaya. There seems, again, no known case of any full Aryan tribe having become savage. The Gypsies and other outcasts are, no doubt, partly Aryan in blood, but their degraded condition is not savagery. In India there are tribes Aryan by language, but whose physique is rather of indigenous type, and whose ancestry is mainly from indigenous stocks with more or less mixture of the dominant Hindu. Some tribes coming under this category, as among the Bhils and Kulis of the

Bombay Presidency, speak dialects which are Hindi in vocabulary at least, whether or not in grammatical structure, and yet the people themselves are lower in culture than some Hinduized nations who have retained their original Dravidian speech, the Tamils for instance. But these all appear to stand at higher stages of civilization than any wild forest tribes of the peninsula who can be reckoned as nearly savages; all such are non-Aryan both in blood and speech.1 In Ceylon, however, we have the remarkable phenomenon of men leading a savage life while speaking an Aryan dialect. This is the wild part of the race of Veddas or 'hunters,' of whom a remnant still inhabit the forest land. These people are dark-skinned and flat-nosed, slight of frame, and very small of skull, and five feet is an average man's height. They are a shy, harmless, simple people, living principally by hunting; they lime birds, take fish by poisoning the water, and are skilful in getting wild honey; they have bows with iron-pointed arrows, which, with their hunting-dogs, are their most valuable possessions. They dwell in caves or bark huts, and their very word for a house is Singhalese for a hollow tree (rukula); a patch of bark was formerly their dress, but now a bit of linen hangs to their waist-cords; their planting of patches of ground is said to be recent. They count on their fingers, and produce fire with the simplest kind of fire-drill twirled by hand. They are most truthful and honest. Their monogamy and conjugal fidelity contrast strongly with the opposite habits of the more civilized Singhalese. A remarkable Vedda marriage custom sanctioned a man's taking his younger (not elder) sister as his wife; sister-marriage existing among the Singhalese, but being confined to the royal family. Mistaken statements have been made as to the Veddas having no religion, no personal names, no language. Their religion, in fact, corresponds with the animism of the ruder tribes of India; some of their names are remarkable as being Hindu,

1 See G. Campbell, Ethnology of India,' in Journ. As. Soc. Bengal, 1866, part ii.

but not in use among the modern Singhalese; their language is a Singhalese dialect. There is no doubt attaching to the usual opinion that the Veddas are in the main descended from the 'yakkos' or demons; i.e. from the indigenous tribes of the island. Legend and language concur to make probable an admixture of Aryan blood accompanying the adoption of Aryan speech, but the evidence of bodily characteristics shows the Vedda race to be principally of indigenous pre-Aryan type.1

The Tatar family of Northern Asia and Europe (Turanian, if the word be used in a restricted sense) displays evidence of quite a different kind. This wide-lying group of tribes and nations has members nearly or quite touching the savage level in ancient and even modern times, such as Ostyaks, Tunguz, Samoyeds, Lapps, while more or less high ranges of culture are represented by Mongols, Turks, and Hungarians. Here, however, it is unquestionable that the rude tribes represent the earlier condition of the Tatar race at large, from which its more mixed and civilized peoples, mostly by adopting the foreign culture of Buddhist, Moslem, and Christian nations, and partly by internal development, are well known to have risen. The ethnology of South-Eastern Asia is somewhat obscure; but if we may classify under one heading the native races of Siam, Burma, &c., the wilder tribes may be considered as representing earlier conditions, for the higher culture of this region is obviously foreign, especially of Buddhist origin. The Malay race is also remarkable for the range of civilization represented by tribes classed as belonging to it. If the wild tribes of the Malayan peninsula and Borneo be compared with the semi-civilized nations of Java and Sumatra, it appears that part of the race survives to represent an early

1 J. Bailey, 'Veddahs,' in Tr. Eth. Soc., vol. ii. p. 278; see vol. iii. p. 70; Knox, 'Historical Relation of Ceylon,' London, 1681, part iii. chap. i. See A. Thomson, 'Osteology of the Veddas,' in Journ. Anthrop. Inst. 1889, vol. xix. p. 125; L. de Zoysa, 'Origin of Veddas,' in Journ. Ceylon Branch Royal Asiatic Soc., vol. vii.; B. F. Hartshorne in Fortnightly Rev., Mar. 1876. [Note to 3rd edition.]

savage state, while part is found in possession of a civilization which the first glance shows to have been mostly borrowed from Hindu and Moslem sources. Some forest tribes of the peninsula seem to be representatives of the Malay race at its lowest level of culture, how far original and how far degraded it is not easy to say. Among them the very rude Orang Sabimba, who have no agriculture and no boats, give a remarkable account of themselves, that they are descendants of shipwrecked Malays from the Bugis country, but were so harassed by pirates that they gave up civilization and cultivation, and vowed not to eat fowls, which betrayed them by their crowing. So they plant nothing, but eat wild fruit and vegetables, and all animals but the fowl. This, if at all founded on fact, is an interesting case of degeneration. But savages usually invent myths to account for peculiar habits, as where, in the same district, the Biduanda Kallang account for their not cultivating the ground by the story that their ancestors vowed not to make plantations. Another rude people of the Malay peninsula are the Jakuns, a simple, kindly race, among whom some trace their pedigree to a pair of white monkeys, while others declare that they are descendants of white men; and indeed there is some ground for supposing these latter to be really of mixed race, for they use a few Portuguese words, and a report exists of some refugees having settled up the country.1 The Melanesians, Papuans, and Australians represent grades of savagery spread each over its own vast area in a comparatively homogeneous way. Lastly, the relations of savagery to higher conditions are remarkable, but obscure, on the American continents. There are several great linguistic families whose members were discovered in a savage state throughout: such are the Esquimaux, Algonquin, and Guarani groups. On the other hand there were three apparently unconnected districts of semi-civilization reaching a high barbaric level, viz., in Mexico and Central America, Bogota, and Peru. Between these higher and

1 Journ. Ind. Archip., vol. i. pp. 295–9; vol. ii. p. 237.

lower conditions were races at the level of the Natchez of Louisiana and the Apalaches of Florida. Linguistic connexion is not unknown between the more advanced peoples and the lower races around them.1 But definite evidence showing the higher culture to have arisen from the lower, or the lower to have fallen from the higher, is scarcely forthcoming. Both operations may in degree have happened.

It is apparent, from such general inspection of this ethnological problem, that it would repay a far closer study than it has as yet received. As the evidence stands at present, it appears that when in any race some branches much excel the rest in culture, this more often happens by elevation than by subsidence. But this elevation is much more apt to be produced by foreign than by native. action. Civilization is a plant much oftener propagated than developed. As regards the lower races, this accords with the results of European intercourse with savage tribes during the last three or four centuries; so far as these tribes have survived the process, they have assimilated more or less of European culture and risen towards the European level, as in Polynesia, South Africa, South America. Another important point becomes manifest from this ethnological survey. The fact that, during so many thousand years of known existence, neither the Aryan nor the Semitic race appears to have thrown off any direct savage offshoot, tells, with some force, against the probability of degradation to the savage level ever happening from high-level civilization.

With regard to the opinions of older writers on early civilization, whether progressionists or degenerationists, it must be borne in mind that the evidence at their disposal

1 For the connexion between the Aztec language and the Sonoran family extending N. W. toward the sources of the Missouri, see Buschmann, 'Spuren der Aztekischen Sprache im Nördlichen Mexico,' &c., in Abh. der Akad. der Wissensch. 1854; Berlin, 1859; also Tr. Eth. Soc., vol. ii. p. 130. For the connexion between the Natchez and Maya languages, see Daniel G. Brinton, in 'American Historical Magazine,' 1867, vol. i. p. 16; and 'Myths of the New World,' p. 28.