2. Professor Rhys has more than once changed his opinion about the Celtic invaders of Britain since he began to handle the subject. In the second edition of his _Lectures on Welsh Philology_[2016] he argued that they were not ‘two distinct nationalities, speaking two distinct languages’; in other words, he maintained that the Goidelic and Brythonic dialects had been evolved within the British Isles after the Celts had entered them. In the preface to _Celtic Britain_, however, which was written in January, 1884, he recanted; and his old view is now obsolete. For many years past he has maintained that the earliest invaders were Goidels, or, as he now prefers to call them, Celticans;[2017] and that the later comers were Brythons. But whereas until a recent date he held that the only Brythonic invasion was that of the Belgae, and that Pytheas, who visited Britain towards the end of the fourth century B.C., ‘is not likely to have found any Brythons here,’[2018] he now holds, or at all events held a few weeks before the time when I am writing, that the first Brythonic invaders ‘appear to have settled here before the middle of the fourth century B.C., for Pytheas ... gives indirect evidence to their presence’.[2019] To this view I hope he will firmly adhere. There is, indeed, no direct evidence that any Brythonic immigrants landed in Britain before the Belgae. But indirect evidence there is; and that of two kinds. The first has been already noticed in the section on the Picts. There are good grounds for believing that the authority whom Diodorus Siculus followed in his notices of Britain was Pytheas.[2020] Diodorus speaks of the British Isles as Πρετ(τ)ανικαὶ νῆσοι;[2021] and the _P_ in Πρετ(τ)ανικαί (if that reading is certain), shows that Pytheas learned the word from lips which spoke a Brythonic, or Gaulish dialect. M. d’Arbois de Jubainville asserts that his informants were Gauls:[2022] but that is simply his opinion; it is open to any one to argue that Pytheas probably learned the name of the Britons as well as the facts which he reported about them and their country in Britain, and not in Gaul. Be this, however, as it may, it is, as we shall presently see, certain that during the earlier period of the Roman occupation, the greater part of England and a considerable part of Scotland were inhabited by Brythons; and, as we shall also see, it is extremely improbable that they were all of Belgic origin. The question of the chronological order of the various Celtic invasions is, according to Professor Rhys,[2023] answered by the present geographical distribution of the Celtic-speaking peoples of the British Isles: ‘it may be regarded,’ he says, ‘as fairly certain that those who are found driven furthest to the west were the earliest comers.’ The argument might be sound enough (though the word ‘driven’ begs the question) if we were considering the British Isles as a whole, and not merely Britain;[2024] and even those who maintain that there were no people of Goidelic descent in Britain in the time of Caesar could hardly answer Professor Rhys unless they assumed that the Goidelic invaders of Ireland came from Spain, or that they dared not risk a contest with the Southern Britons; for otherwise it is hard to believe that they would not have directed their immigration towards Britain, the nearer country.