[[NB. Footnote numbers (in square brackets) differ from those in the print edition, where a new series begins on each page. References to ancient texts are given here in modern form. Lower case Roman numerals (in round brackets) indicate the page number in the print edition.]]
That the son-in-law of Agricola was in possession of a very considerable amount of first-hand authentic information concerning Agricola's family, training, and early official career goes without saying. And for the events and data subsequent to Agricola's departure for Britain, Tacitus doubtless also learned innumerable details from the living lips of Agricola himself, although he is cited directly as a source but three times (chs. 4.3. 24.3, and 44.5). Again, many items of information would naturally come to his knowledge as being a contemporary; still others were probably transmitted to him by some who were with Agricola in Britain. For the narrative of the years 89-93, during which the author was absent from Rome, he could also rely upon trustworthy oral testimony furnished to him on his return.
In fact, the only items of information which involve the assumption of a distinctly literary source are contained in the survey of the earlier predecessors of Agricola and their military campaigns. What this was cannot be determined. Perhaps Fabius Rusticus was one of them, but that detailed narratives were at his disposal is shown by the full account given of some of these events, such as the uprising under Boudicca (ch. 15f.), in his Annals.
The description of the battle at Mt. Graupius, barring a certain historical background, and particularly the speeches (xxiii) were, as we have seen, largely, if not wholly, works of the creative imagination.
To what extent, finally, books, if any, were consulted for the account of the ethnology and geography of Britain, it is impossible to determine. Tacitus himself was certainly acquainted with the most important works dealing with these topics, for he singles out Livy and Fabius Rusticus as the most famous out of many accessible to him, assuring us that their information was scanty and unreliable, he being in possession of more authentic details. Caesar, whom Tacitus styles summus auctorum in the Germania (ch. 28.1), seems not to have been consulted, or rather he was intentionally ignored, for he cannot be included among those who nondum comperta eloquentia percoluere (ch. 10.1), a statement in no sense applicable to the short paragraphs which Caesar has devoted to the subject.
That he made direct use of any Greek sources, such as Pytheas of Massilia, Strabo, Diodorus, or even Posidonius, whose works on physical geography and ethnology marked the highest point which scientific research into these subjects reached in antiquity, is extremely doubtful, for Tacitus's explanations of the long days and the tides are unscientific; and in holding to the disk-shaped form of the earth, he is deplorably behind the knowledge even of his own time.
 See the exhaustive treatment of this subject in Furneaux's Introd. to the Agricola, pp. 22-34.
 See Agr. 10.3.
 BG 4.33; 5.12-14. Caesar seems to imply that the Britons had cavalry as well as chariot fighters, whereas Tacitus mentions only the latter, whom he calls covinnarii, for which Caesar says essedarii. But the contradiction is, perhaps, only apparent, for Tacitus is speaking of the Caledonians, Caesar of the Britons in Kent, and what was true of the latter was not necessarily applicable to the former.