us to give precision and point to much that troubles students in obscurer texts. When Mr. Fitzgerald talks about Crowe's translation of the Amra Choluim Cille, we have to decide what 'translation' means. He may be quite sure that the interests of Celtic story, mythological, or other, are not being neglected during this long and arid process. Many a tough conflict of wordy criticism will have to be fought. A good deal of heat will have to be generated before the goal is attained. The battle has not been without wounds and clamour so far, nor is there any reason to suppose that Irish can be freer than other philology from the baleful results of literary vendetta. The truth is, that Celtic studies labour under the fatal disadvantage of having no competent public to oversee and control: the criticism is frequently harsher than the occasion deserves. But there is unmistakeable progress; enlightenment is spreading; the knowledge once possessed by the privileged few is now of comparatively easy attainment; and the leaven of an eager desire to get to the bottom of these Celtic mysteries is working in many minds. The difficulties are mainly lexicographical: we do not know the import of many a word that occurs in our Irish texts, and we are not permitted the free license of guessing indulged in by the past generation. Mr. Fitzgerald might remember that there is not a page of his article that does not contain speculations or references depending on the meaning of individual words. Take, e.gr., the following passage, p. 201: "This collocation or opposition of the Hound and the Dog reappears in the ancient division of Ireland into Conn's Half (the North), and Mog Nuadat's Half (the South): for Mog Nuadat is but the Servant-of-the-Hand … and there is reason to suspect that in effect the other name is the Dog's (or Wolf's) Half." Now this is a speculation based on the meaning of Nuadat and of Conn; and by what means can the settlement of the question be sought other than by the comparison of texts, in which these words are used in the meanings alleged? Words, words, words, that is what we want, and for many a long year they will be wanted. At present scarcely the simplest piece of Middle Irish prose of a few pages long but contains some word or phrase that must be passed over with a query or a blank space. It is useless arguing that Nuadat means hand, unless texts can be produced clearly establishing this meaning, which certainly has not been done yet. But it is astonishing how soon the note of interrogation drops away from a hypothesis.
