This paper is a detailed study of the language of the writing-tablets from Vindolanda reported on by Bowman et al. in this volume.

Professor Michael Speidel has recently suggested that the abbreviation SC on military inscriptions does not, as is commonly thought, refer to singularis consularis but to summus curator. In some instances his interpretation is without doubt correct. However, his view that the five examples from Britain are not the guards of the provincial governor but chief accounts-clerks needs further consideration. Speidel argues that the post of singularis consularis is usually abbreviated to sing. or singul. cos. This is quite true, but it must be noted that examination of all the different examples of singulares shows that the post could be abbreviated in a large number of ways; even in the same document it may be written and abbreviated in different ways by the same clerk. At Apulum, the provincial capital of Dacia, no less than ten different ways are used on tile-stamps to refer to the governor's guards, while inscriptions there reveal a further three. Among the tile-stamps is the abbreviation SC, here without any doubt meaning singulares consularis; consequently, there is a firm Continental parallel for the traditional interpretation of the British examples, which Speidel would have changed. Presumably the British (and African) troops preferred a particular style of abbreviation, which was not so favoured elsewhere in the Empire. The only soldier in Britain who was without doubt a strator co(n)s(ularis) significantly and deliberately chose to write the first part of his title out in full, although this meant abbreviating and ligaturing his own name; this suggests that he wished to avoid confusion over the meaning of the normal abbreviation SC, which—in Britain at any rate—referred to singularis consularis.

Although the so-called vexillation fortress at Osmanthorpe (SK 678563) in the Greet Valley, Nottinghamshire, is a recent discovery, fieldwork in the immediate vicinity of the site has been conducted independent of it in the past. This paper reports the results of recent work connected with what we shall term the castra and brings to wider attention some of the earlier work in the Greet Valley, combined with a consideration of its possible implications.

Recent excavation and coring of the collapsed east wall of the Saxon Shore fort of Richborough has revealed the manner in which the wall collapsed. This led to a re-evaluation of the original siting of the wall, which must have lain to the west of where it is usually depicted. Reassessment of previous excavations, including the examination of original records from the J.P. Bushe-Fox excavations of the 1920s and 1930s leads to the conclusion that the so-called ‘unfinished’ or ‘abandoned’ east wall foundation was in fact the base of the built east wall, from which the collapse derives. A revised fort plan based upon this conclusion is suggested. Supplementary material is available online (https://doi.org/10.1017/S0068113X20000379), and includes additional backing tables and illustrations referenced in the text.

An alternative is proposed to the interpretation of the plan of Dalswinton Roman fort offered by W.S. Hanson and colleagues in their important study of the air-photographic, geophysical and LiDAR evidence published in Britannia in 2019. It is suggested that the larger, second-phase fort faced not east, but south, as previously thought, but that most of the remains of this layout have been removed by the plough. A suggestion is made concerning the possible garrison of the second-phase fort.

An unusual copper-alloy fragment was recovered during excavations at Vindolanda in 2008. It has been identified as part of a calendar or water clock. A very similar fragment was found near Hambledon in Hampshire in 2017. Further investigation of the Vindolanda and Hambledon fragments and of similar objects from Frankfurt, Salzburg and Grand reveals that the Vindolanda and Hambledon fragments were once attached to clepsydrae to form time-keeping devices that are unattested in ancient sources, but that might be called ‘calendrical clepsydrae’.1 The links between similar artefacts and deities associated with water and healing are explored and evidence for a shrine located near sulphur springs in the Allen Valley is also discussed. The Vindolanda calendrical clepsydra is placed within this social and religious context.

The original poem inscribed above an image of Europa and the bull on the Lullingstone villa mosaic has long been recognised as referring to an episode in Ovid's Metamorphoses and making an allusion to Virgil's Aeneid. Here it is proposed that it is also a parody of a poem by Martial on the same subject and adds to the growing quantity of classical literature with which the fourth-century Romano-British elite are known to have been familiar.

Cae Metta (SH 536650), 1.5 km south-west of Llandeiniolen: preliminary examination of the polygonal enclosed hut-group suggested occupation from the mid third into the fourth century.

Caernarvon fort (Segontium, SH 485624) :3 excavation outside the north-east gate (porta decumana) showed that the original Flavian causeway and inner ditch remained in use unaltered until shortly after 330. The causeway was then dug through and the ditch re-cut from a W-profile to the normal V-section, while a new outer ditch was constructed some 10 m out from the gate. Access to the gate was now over wooden bridges, for which structural evidence was found at the inner ditch. Fresh details were also obtained of the Flavian timber gate.