fragments of the literal originals only, were discovered. The same argument, however, applies with much stronger force to such a collection of ancient ballads as Scott's Minstrelsy of the Scottish Border, which does not suffer the overwhelming disadvantage of being a translation. It is certain that no literal copy of any one of the ballads collected and printed by Sir Walter Scott could have been found by Yarrow or among the Cheviot hills. The printed versions, as is well known, were collated by Scott from the recitals of various persons, and bear only part resemblance to the tradition of each. In the same way Macpherson collated the various copies of ancient poems which he managed to procure. Such gaps in the poems as occur, for instance, in "Colna-dona," and the first duan of "Cath-loda," offer testimony to his style of work. This fact, together with the difficulty translation inevitably introduces, besides the altering of titles and phoneticising of names which Macpherson effected, was quite sufficient to preclude the possibility of any of his transcripts being found complete and exact in the original. The Report of the Highland Society alludes to this circumstance. It says (p. 152), "The advantage he (Macpherson) possessed, which the Committee began its inquiries too late to enjoy, of collecting from the oral recitation of a number of persons, now no
