In the centre of the island is the kingdom of Kandy; naturally fortified by impervious forests, and long independent. This creates a variety; the Kandyans being somewhat ruder than the other Singhalese. It is not, however, an important one. The really important ethnology of Ceylon is that of the _Vaddahs_, in the eastern districts, inland of Battacaloa. They are still unmodified by either the Hindú habits, or the great Indian creeds,--the true analogues of the Khonds, and Kóls, and Bhils, &c. Their language, however, is Singhalese; an important fact, since it denotes one of two phenomena,--either the antiquity of the conquest of Ceylon supposing the extension of the Singhalese language to have been gradual, or the thorough-going character of it, if it be recent. Who were the _Padæi_ of the following extract from Herodotus?[56]--"Other Indians there are, who live east of these. They are nomads, eaters of raw flesh; and called Padæi. They are said to have the following customs. Whenever one of their countrymen is sick, whether man or woman, he is killed. The males kill the males, and amongst these the most intimate acquaintance kill their nearest friends; for they say that for a man to be wasted by disease is for their own meat to be spoilt. The man denies that he ails; but they, not letting him have his own way, kill and feast on him. If a female be sick, the women that are most intimate with her treat her as the males do the men. They sacrifice and feast upon all who arrive at old age. Few, however, go thus far, since they kill every one who falls sick before he reaches that stage of life." Name for name, the _Vaddahs_ of Ceylon have a claim to be _Padæi_. Besides which they are Indian. But, name for name, the _Battas_[57] of Sumatra have a claim as well; and although they are not exactly Indian, they are cannibals of the sort in question--or, at any rate, cannibals in a manner quite as remarkable. This gives us a conflict of difficulties. The solution of them lies in the fact of neither _Vaddah_ nor _Batta_ being _native_ names; a fact which leaves us a liberty to suppose that the _Padæi_ of Herodotus were simply some wild Indian tribe sufficiently allied in manners to the _Vaddahs_ of Ceylon, and the _Battas_ of Sumatra, to be called by the same name, but without being necessarily either the one or the other; or even ethnologically connected with either. * * * * *