Recent volumes of the Philological and Historical Commentary on Ammianus Marcellinus, for many years now the joint effort of a Quadriga Batavorum, have been superb. In breadth of vision and quality of commentary they have moved far beyond the original plans of P. de Jonge, who began the project as long ago as 1935. De Jonge passed the torch to three of the present authors in the mid-1980s with Book 20, and Drijvers joined the team for the publication of the commentary on Book 22 in 1995. With the publication in rapid succession of Books 29 (the volume under review) and 30 (2015), the whole towering monument is now approaching completion. Only the monographic Book 31 remains, with its long and evocative tale of the Gothic war and the trauma of Adrianople.

Commentaries only reveal their true value with use. Reading one cover to cover is at the very least fatiguing and can serve to dampen more than fire one's enthusiasm for the author in question; the Dutch commentators’ Ammianus volumes are no exception to that rule. Yet that having been said, Ammianus’ Books 26–30 benefit more than their predecessors from sustained commentary. They function as a coherent unit of post-Julianic narrative, but in terms of their content, and even more the clarity of their prose, they are not always the equal of the earlier books – they pose real difficulties of both historical and philological interpretation. For several years now, G. Kelly and I have been working on an annotated translation of the whole of Ammianus, to replace Rolfe's antiquated Loeb and Hamilton's sadly abridged Penguin, for Anglophone audiences. With the above criterion of actual use thus satisfied, it seems safe to say that the Quadriga’s commentaries, on the five late books in particular, are an irreplaceable resource. Book 29 is one of the most rebarbative in the whole of the Res Gestae. While it begins with the famous account of the divination and treason trials under Valens in Antioch – which took place while Ammianus was resident in the city – the rest of the book skitters along without much of a narrative arc. It also includes in its enormous fifth chapter, the main (virtually the only) source for the rebellion of the Moorish nobleman Firmus. Modelled to some extent on Sallust's Jugurtha, this chapter is perhaps the single most difficult portion of the entire Ammianean narrative, its geography obscure and its chronology extremely uncertain.

The commentators’ treatment of the treason trials (1–2) is exemplary and particularly good at spotting the implications of rare words or Ammianus’ deliberate ambiguities (for instance his fidenter at 19.1.6, to imply a witness’ being offered immunity in exchange for testimony). However, this section also makes absolutely clear how important it is to have all the commentary volumes at hand when working with any of them: in several crucial places, one must refer back to an earlier volume for a full discussion of, for example, lese-majesty (covered in the commentary on 21) and the details of a torturer's equipment (in 26). Chapters 3 and 4 focus on Valentinian and affairs in the West, and again there are important insights. The discussion of Ammianus’ use of subagrestis for the Pannonian emperors (8 of 11 CLCT citations of the adjective are from Ammianus!) is particularly helpful in drawing out both Ammianus’ snobbery and his predilection for singling out certain regional backgrounds as flawed or undesirable. The famous pair of savage bears, Mica Aurea and Innocentia, which Valentinian is said to have kept to have enemies thrown to (3.9), are present and correct, though why the commentators translate the first bear's name as ‘Golden Kid’ rather than ‘Goldflake’ is unclear – mica connotes something small, or wispy or slight, not something young. More broadly, Valentinian's bears are another place where, as is generally their way, the authors simply decline to pass judgement on the authenticity of an Ammianean anecdote, although some scholars reject it as wholly false and others accept it entirely.

The most important part of Book 29 remains the fifth chapter and the commentators’ treatment can now be regarded as the essential point of reference for future discussions. At point after point, they canvass diverse interpretations and arrive at a judicious conclusion. They are surely right, for instance, that the Nubel whom Ammianus describes as a regulus and the founder of Firmus’ dynasty is identical with the Flavius Nuvel of CIL VIII 9255, thus the praepositus of a cavalry detachment whose father was an imperial comes named Saturninus. That identification is significant because it means that ‘Nubel belonged to both the Roman and the Moorish culture’ (p. 153) and that we must not improperly exoticise him as an indigenous rebel. In fact, the commentators might have gone further in pointing out how typical of the later fourth-century empire Nubel and his family actually were: parlaying local status into a successful military or administrative career at the imperial level was a classic provincial manoeuvre which, if handled right, would in turn entrench a pre-existing local status. Nubel's clan has looked exotic to modern scholars because of their Moorish nomenclature, but apart from their background in sparsely urbanised Mauretania, their social location was not much different from that of the elder Gratian or the elder Theodosius, and perhaps somewhat higher than that of, say, Aurelius Victor or Ammianus himself. Nevertheless, the basic clarity of the discussion here is excellent, and the commentators are correct to limit the scene of Firmus’ revolt to the eastern part of Mauretania Caesariensis and not exaggerate it into a rebellion across swathes of both Mauretanias. Likewise, the dating of various stages of the Firmus revolt is notoriously difficult and probably impervious to definitive resolution, but here every piece of chronological evidence internal to the discussion (and really there are almost no external clues outside of Ammianus’ own text) is handled fairly. More importantly, the end of the revolt and Theodosius’ triumphant return to Sitifis is correctly placed in the year 375, before the death of Valentinian I, but definitely quite some time after the last firm date in the text (Theodosius is at Tifasa in February 374: 5.31).

This volume is a very significant achievement, as too is the completion of the larger commentary on Books 26–30. Taken together, they render these less shapely portions of the Res Gestae almost as legible as the Julianic books and, in this volume, bring a great deal of good sense to one of the most difficult chapters in the whole Ammianean corpus.