After long neglect, in English-language scholarship at least, the question of how Julius Caesar wrote and disseminated hisGallic War—as a single work? in multi-year chunks? year by year?—was revived by T.P. Wiseman in 1998, who argued anew for serial composition. This paper endeavours to provide further evidence for that conclusion by examining how Caesar depicts the non-Roman peoples he fights. Caesar's ethnographic passages, and their authorship, have been a point of contention among German scholars for over a century, but reading (...) them and the rest of the text with eyes unclouded by the exhausted debate about possible interpolation reveals details that bear upon wider questions of composition. In these passages Caesar devised an ethnographic framework in order to rank against one another the levels of threat posed by different barbarian peoples, downplaying the relative ferocity of the Gauls in contrast to other groups in an effort to magnify the peril the others posed to Rome and the glory to be obtained from their defeat. This ethnographic framework is significant for understanding Caesar's method both because it provides insight into Caesar's reasons for including the ethnographic passages and because it implies that theGallic Warwas composed in, at a minimum, four stages: Books 1–2, where the framework is first developed and used, by 56b.c.; Books 3–4 and 5–6, where it is elaborated and extended, by 54 and 52b.c.respectively; and finally Book 7, after 52b.c., when Caesar, in recounting the campaign against Vercingetorix, was forced to abandon and contradict the ethnographic framework in a fashion that suggests that the earlier books were already in circulation, preventing him from adjusting them to the new circumstances of the campaign of that year. (shrink)