cated animals of civilization, have brought in with them a corresponding art and knowledge. The course of events which carried horses and wheat to America carried with them the use of the gun and the iron hatchet, while in return the whole world received not only maize, potatoes, and turkeys, but the habit of tobacco-smoking and the sailor's hammock.
It is a matter worthy of consideration, that the accounts of similar phenomena of culture, recurring in different parts of the world, actually supply incidental proof of their own authenticity. Some years since, a question which brings out this point was put to me by a great historian — 'How can a statement as to customs, myths, beliefs, &c., of a savage tribe be treated as evidence where it depends on the testimony of some traveller or missionary, who may be a superficial observer, more or less ignorant of the native language, a careless retailer of unsifted talk, a man prejudiced or even wilfully deceitful?' This question is, indeed, one which every ethnographer ought to keep clearly and constantly before his mind. Of course he is bound to use his best judgment as to the trustworthiness of all authors he quotes, and if possible to obtain several accounts to certify each point in each locality. But it is over and above these measures of precaution that the test of recurrence comes in. If two independent visitors to different countries, say a mediæval Mohammedan in Tartary and a modern Englishman in Dahome, or a Jesuit missionary in Brazil and a Wesleyan in the Fiji Islands, agree in describing some analogous art or rite or myth among the people they have visited, it becomes difficult or impossible to set down such correspondence to accident or wilful fraud. A story by a bushranger in Australia may, perhaps, be objected to as a mistake or an invention, but did a Methodist minister in Guinea conspire with him to cheat the public by telling the same story there? The possibility of intentional or unintentional mystification is often barred by such a state of things as that a similar statement is made in two remote