his edition of Sir Tristrem, published in 1804, and that continued to interest him vividly as long as he lived.  It reappears in many of his critical writings[78] and also in the novels.  In the Bride of Lammermoor Ravenswood goes to his death in compliance with the prophecy of Thomas quoted by the superstitious Caleb Balderstone.  And in Castle Dangerous Bertram, who is unconvincing perhaps because he is endowed with the literary and antiquarian tastes of a Walter Scott himself, is actuated by an irrepressible desire to discover works of the Rhymer.

Scott’s edition of Sir Tristrem gives—­besides the text, introduction, and notes—­a short conclusion written by himself in imitation of the original poet’s style.  Much of his theory has fallen.  He considered this Sir Tristrem to be the first of the written versions of that story, a supposition that was not long tenable.  The poem is now known to be based upon a French original, and many scholars think the name Erceldoune was arbitrarily inserted by the English translator; though Mr. McNeill, the latest editor, thinks there is a “reasonable probability” in favor of Scott’s opinion that the author was the historic Thomas, who flourished in the thirteenth century.  It is important, however, that Scott’s scholarship in the matter passed muster at that time with such men as Ellis, who wrote the review in the Edinburgh, in which he said, “Upon the whole we are much disposed to adopt the general inferences drawn by Mr. Scott from his authorities, and have great pleasure in bearing testimony to the very uncommon diligence which he has evinced in collecting curious materials, and to the taste and sagacity with which he has employed them....  With regard to the notes, they contain an almost infinite variety of curious information, which had been hitherto unknown or unnoticed."[79] John Hookham Frere said, as quoted in a letter by Ellis, “I consider Sir Tristrem as by far the most interesting work that has as yet been published on the subject of our earliest poets."[80] Scott’s opinions were in 1824 thought to be of sufficient importance, either from their own merits or on account of his later fame, to call forth a dissertation appended to the edition of Warton’s History of English Poetry published in that year.

The first edition of the text swarms with errors, according to Koelbing,[81] a recent editor of the romance, and later editions are still very inaccurate.[82] It could hardly be expected that a man with Scott’s habits of mind would edit a text accurately.  But no one of that period was competent to construct a text that would seem satisfactory now.  The study of English philology was not sufficiently developed in that direction, nor did scholars appreciate either the difficulties or the requirements of text-criticism.  It is not to be wondered at that Scott failed, in this instance as well as afterwards in the case of the text of Dryden, to give a version that would stand the minute scrutiny of later scholarship.