From a palæographic point of view the oldest books in Ireland are probably the "Domhnach Airgid," a copy of the Four Gospels in a triple shrine of yew, silver-plated copper, and gold-plated silver, which St. Patrick was believed to have given to St. Carthainn when he told that saint with a shrewd wisdom, which in later days aroused the admiration of Mr. Matthew Arnold, to build himself a church "that should not be too near to himself for familiarity nor too far from himself for intercourse." It probably dates from the fifth or sixth century. The Cathach supposed to have been surreptitiously written by Columcille from Finnian's book[8]--a Latin copy of the Gospels in Trinity College, Dublin; the Book of Durrow, a beautiful illuminated copy of the same; the Book of Dimma, containing the Four Gospels, ritual, and prayers, probably a work of the seventh century; the Book of Molling, of probably about the same date; the Gospels of Mac Regol, the largest of the Old Irish Gospel books, highly but not elegantly coloured, with an interlinear Anglo-Saxon version in a late hand carried through its pages; the Book of Kells, the unapproachable glory of Irish illumination, and some other ecclesiastical books. After them come the Leabhar na h-Uidre and the great books of poems and saga. Although the language of these sagas and poems is not that of the glosses, but what is called "Middle-Irish," still it does not in the least follow that the poems and sagas belong to the Middle-Irish period. "The old Middle-Irish manuscripts," says Zimmer, "contain for the most part only Old Irish texts re-written."[9] "Unfortunately," says Windisch, "every new copyist has given to the text more or less of the linguistic garb of his own day, so that as far as the language of Irish texts goes, it depends principally upon the age of the manuscript that contains them."[10] And again, in his preface to Adamnan's vision, he writes: "Since we know that Irish texts were rewritten by every fresh copyist more or less regularly in the speech of his own day, the real age or a prose text cannot possibly be determined by the linguistic forms of its language."[11] It is much easier to tell the age of poetry than prose, for the gradual modification of language, altering of words, shortening of inflexions, and so on, must interfere with the metre, so that when we find a poem in a twelfth-century manuscript written in Middle Irish and in a perfect metrical form, we may--no matter to what age it is _ascribed_--be pretty sure that it cannot be more than two or three centuries older than the manuscript that contains it. Yet even of the poems Dr. Atkinson writes: "The poem _may_ be of the eighth century, but the forms are in the main of the twelfth."[12] Where poems that really are of ancient date have had their language modified in transcription so as to render them intelligible, the metre is bound to suffer, and this lends us a criterion whereby to gauge the age of