Other matters that necessarily presented themselves when he was treating the subject of romance were the problem of the sources of narrative material, especially the perplexed question concerning the development of the Arthurian cycle, and the problem, already discussed in connection with ballads, concerning the character of minstrels. The minstrels reappear throughout Scott's studies in mediaeval literature, and were perhaps more interesting to him than any other part of the subject. Though, as we have seen, he formulated a compromise between the opposing opinions of Percy and Ritson, no one who reads the description of the Last Minstrel can doubt what was the picture that he preferred to carry in his mind. His ideas on the subject of the origin and diffusion of narrative material were those of the sensible man trying to look at the matter in a reasonable way. Here again he adopted an attitude of compromise, in that he admitted the partial truth of various theories which he considered erroneous only in so far as any one of them was stretched beyond its proper compass. "Romance," he said, "was like a compound metal, derived from various mines, and in the different specimens of which one metal or other was alternately predominant."[86] On the subject of the Arthurian cycle, the origin of which has never ceased to be matter for debate, he held essentially the opinions that the highest French authority has adopted that Celtic traditions were the foundation, and that the metrical romances preceded those in prose.[87] The important offices of French poets in giving form to the story he underestimated. When he said, "It is now completely proved, that the earliest and best French romances were composed for the meridian of the English court,"[88] he fell into the error that has not always been avoided by scholars who have since written on the subject, of feeling certitude about a proposition in which there is no certainty. Scott's work on romances, though it does not always rise above commonplaceness, escapes the perfunctory quality of hack writing by virtue of his keen interest in the subject. He continued to like this prosaic kind of literary task even while he was writing novels with the most wonderful facility. We may judge not only by the fact that he continued to write reviews at intervals throughout his life, but by an explicit reference in his _Journal_: "I toiled manfully at the review till two o'clock, commencing at seven. I fear it will be uninteresting, but I like the muddling work of antiquities, and besides wish to record my sentiments with regard to the Gothic question."[89]