islands of the Antilles; the aboriginal inhabitants of Brazil; the Abipones, so well described by Dobrizhofer (1784); and in the South, the Patagonians and the inhabitants of Tierra del Fuego.

Until the languages of these people have been carefully analysed by real scholars, any attempt at grouping them would prove simply mischievous. We are at present in a stage where our duty is to distinguish, not to confound. Even to speak of the inhabitants of Tierra del Fuego as one race has produced, as we saw, disastrous results, and it is to be hoped that we shall hear no more of a South American language or of a North American religion. It is true that certain legends have been found in the North as well as in the South of America, which seem to point to a common origin. But it will be time to account for such coincidences after the legends of each centre have been studied by themselves, and after some clearer light has been thrown on the component elements of the population of the whole American continent.

How, under present circumstances, scholars could have been bold enough to trace the whole American race to immigrations from Asia or even from Europe, is difficult to understand. The physical possibility, no doubt, was there, whether across the island bridges in the North, or by sea from West or East. We heard but lately how a large vessel, cast off by its crew, drifted safely from America to England (the Hebrides). The same may have happened on either coast of America. But any attempts to recognise in the inhabitants of America descendants of Jews, Phenicians, Chinese, or Celts are for the present

simply hopeless, and are in fact outside the pale of real science.

Oceanic Languages.

The languages which extend from Madagascar on the East coast of Africa to the Sandwich Islands, West of America, have been far more carefully studied than those of America and Africa. I speak of languages, not of races, for if ethnological classification has proved a failure anywhere, it has when applied to the mixture of blood that led to the formation of such races as Australians, Papuans, Malays, Polynesians, Melanesians, Micronesians, Negritos, Mincopies, Orang-utans, and all the rest.

From the latest work on this family of languages, by Dr. Codrington (The Melanesian Languages,' Oxford, 1885), it appears that we must admit an original, though very distant, relationship between the Malay, the Polynesian, Melanesian, and Micronesian languages, but that in their later development it is possible to distinguish between the Malay, the Polynesian, and the Melanesian (with Micronesian) as independent branches of a common stem. The dialects of Australia stand as yet apart, as too little known, as well as those of New Guinea, though some dialects, like the Motu of New Guinea, are clearly Melanesian.

It follows from this division, that with regard to religion also we must distinguish between a Malay, a Polynesian, a Melanesian, and possibly a New Guinea (Papuan) and Australian centre. Our information, however, from the two last, is very imperfect.

Malay.

Owing to the proximity of the Malay islands to India, they have from the earliest times been overrun by immigrants, conquerors, and missionaries from the Asiatic Continent. Their ancient ancient religious opinions are covered up and hidden under superimposed strata of Hindu, Buddhist, Mohammedan, and Christian faith, and what there is of native growth in Java, Borneo and elsewhere represents probably the mere dregs of a former religion.

Polynesia.

The Polynesian languages, on the contrary, present us with an abundant growth both of religion and of poetical mythology. These Polynesian traditions are particularly valuable to the student of comparative mythology, because they offer striking similarities with the legends of Greeks, Romans, Teutons and others, without the possibility of a common origin or of a later historical contact.

Melanesia.

The Melanesians, so far as we can judge, do not differ much from the Polynesians and Micronesians in the fundamental outlines of their religious opinions, but they are not so rich in imaginative legends. Further research, however, may modify this opinion.

As to the Australians and the Papuas of New Guinea, very little has been ascertained as yet of their religion, except what is embodied in their ceremonial observances and social customs.

Classification of Languages, why necessary.

This linguistic and religious survey, which has

taken up much of our time, will nevertheless, I hope, prove a saving of time in the progress of our work. Imperfect as it is, it will enable us to guard against certain mistakes very common in the Science of Religion. We have established certain broad lines of division in language and religion, and we shall hear no more of what used to be called the religion of savages, or barbarians, or black men, or red men, or Africans, or Americans. The student of religion knows no savages, no barbarians. Some of the races who are called savage or barbarous possess the purest, simplest, and truest views of religion, while some nations who consider themselves in the very van of civilisation, profess religious dogmas of the most degraded and degrading character. The African Zulu who was a match for Bishop Colenso, cannot be classed as an African or black man together with the royal butchers of Dahomey; and the Inca philosopher who searched for something more divine than the sun, cannot be placed by the side of the Blackfoot performing the sun-dance 1.

Progress in the Science of Religion means at present discrimination, both with regard to the subject and the object of religious faith. As we speak no longer of the believers in a religion as either savages or barbarians, black men or red men, Africans or Americans, the idea also that we can truly characterise any religion by such general terms as fetishism, totemism, animism, solarism, shamanism, etc., has long been surrendered by all critical students. In

The Blackfoot Sun-Dance, by Rev. John McLean, in the Proceedings of the Canadian Institute, No. 151; 1889. Notes bearing on the use of ordure in rites of a religious character, by John G. Bourke, Washington, 1888.

gredients of all these isms may be found in most religions, but not one of them can be fully defined by such vague terms. Religions are everywhere the result of a long historical growth, and, like languages, they retain even in their latest forms traces of the stages through which they have passed. There is fetishism in some forms of Christianity; there is spiritualism in the creed of some so-called worshippers of fetishes. Generalisation will come in time, but generalisation without a thorough knowledge of particulars is the ruin of all sciences, and has hitherto proved the greatest danger to the Science of Religion.