Since the 1970's large scale work at Strageath (Frere and Wilkes 1989) has largely solved that site's chronological questions, but the passage of almost thirty years has left Bertha and Dalginross almost as mysterious as ever, despite the publication of excavations from both sites (Adamson and Gallagher 1986 and Robertson 1964). The Cumbernauld Society's remaining finds were thus still potentially a very precious resource and in 1998 the Society was kind enough to lend them to the Roman Gask Project1, prior to donating them to Perth Museum. The finds were analysed by F.C. Wild, A.T. Croom (Tyne and Wear Museums) and K.F. Hartley, to all of whom the Gask Project is most grateful.

Results

Strageath
This fort provided much the smallest corpus of finds from the Cumbernauld Society's field walking program, a total of only eight sherds, and no fine wares were recovered. Nevertheless, the three mortarium fragments present have added some small confirmation to the picture of both Flavian and Antonine occupation already established by the much larger finds body from the Frere and Wilkes (1989, 117-138) excavations.

Bertha
As one of the three Gask system forts, it has usually been assumed that Bertha would have followed the same sequence as its sister forts at Ardoch and Strageath, both of which show Flavian and Antonine occupation. Excavation has now confirmed Flavian activity (Adamson and Gallagher 1986), but the only evidence for later occupation has remained an inscription to Discipulinae Augusti (Keppie 1983, 402) recovered from the Tay beside the fort, a type which tends to be 2nd century or later in date and thus seemed more likely to be Antonine, or even Severan, than Flavian. Antonine occupation now seems to have been firmly substantiated by the Cumbernauld Society's finds, since all of the fine wares and mortaria fragments have been designated below as either "Antonine" or "probably Antonine". The material has thus provided welcome confirmation for an aspect of our existing model of the history of Roman Scotland. The fort also yielded five fragments of Medieval pottery suggesting later occupation on or near the site.

Dalginross
This site is a member of the line of so called "Glen blocking" forts which run along the southern highland fringe from Drumquhassle to Stracathro and which include the legionary fortress of Inchtuthil. This line has always been regarded as a probably fairly short lived and purely Flavian system for, when excavated, its installations have rarely shown more than a single structural phase. Moreover, with the exception of a single coin of Trajan (Scots Magazine 1771, 501 and Macdonald 1918, 245) found on or close to the fort of Drumquhassle and a coin of Severus Alexander which may have been found near to (but not in) Dalginross (Macdonald 1924, 326), no finds of any other period have ever been recovered from these forts. Dalginross has long been one of the worst dated of the line but, even here, the position seemed assured for, although no finds whatever were recovered from the only excavation to have been conducted on the site (Robertson 1964, 198), two stray coin finds (Macdonald 1924,326), an aureus of Titus and an As of Domitian (dated to 86 A.D.), appeared to confirm the Flavian date. Here, however, the Cumbernauld finds are able to offer their biggest surprise and, arguably their greatest contribution to our understanding of the Roman occupations of Scotland, for the site was found to have yielded both Flavian and Antonine material. Indeed, only one of the five Samian sherds recovered can be assigned to the Flavian period (see below) and all three of the mortarium fragments appear to be Antonine.

The Samian from Dalginross had not been marked with field reference numbers and so unexpected were the Antonine identifications that on receiving the fine ware report (the first to be completed), the writer contacted the Cumbernauld Society to ask if there was any possibility that the material could have been mixed up since its discovery. The answer was that the finds had once been taken to show a school class and that although careful precautions had been taken, it was just possible that material from one fort could have been replaced in the wrong finds bags. This appeared to put some doubt on the Samian analysis, but the coarse ware fragments had been marked in the field and these were certainly all found to be in the correct bags. This means, at the very least, that the coarse ware reports can be completely relied upon and, as the material was all bagged together in one bag per fort, it might also suggest that none of the material had been mixed, which would allow us to regain rather more confidence in the Samian datings. Whatever the case, however, the mortaria datings alone are sufficient to provide a firm indicator of Antonine occupation at the site, which means that we are forced to re-evaluate our picture of Antonine activity in northern Scotland. As has already been said, the Gask forts of Ardoch and Strageath were already known to have been re-occupied in the Antonine period, presumably as outposts to the Antonine Wall, and the Cumbernauld material has now confirmed the long held assumption that Bertha would also have fitted this pattern (the position at Doune remains unclear). It would now seem that we must also add Dalginross to the list and this may have wider repercussions.

Firstly, with the (possible) exceptions of Inchtuthil and Fendoch, the entire glenblocking line has been almost uniformly understudied, especially by excavations using modern techniques. Indeed some sites, such as Malling, remain virtually untouched. For many years we have simply felt that the dating issues were secure and that there was, thus, little point in devoting additional effort to a series of sites which were not (Drumquhassle to some extent excepted) under threat. It would now seem that this confidence may have been misplaced, however, and the possibility arises that other glenblocking forts might also have seen Antonine activity. This might appear at least somewhat unlikely because Dalginross has produced what might now be seen as an important feature which is, so far, unique on the line: two concentric fort shaped enclosures (Robertson 1964 199, fig 9). It has usually been suggested that the inner of these represented the fort and the outer, an annex or construction camp (e.g. Robertson 1964, 198, although cf St Joseph 1951, 64) but it is possible that they are simply the defences of the Flavian and Antonine forts (although the exact Cumbernauld find spots were not recorded making it impossible to determine which is which). Nevertheless, Drumquhassle is also completely encircled by an, albeit less regular, enclosure, which has also been generally regarded as an annex (Maxwell 1983, 168-172 and fig 2), and it might now seem desirable to conduct further work on the glenblockers, either excavation or, at the very least, additional field walking.

Secondly, Dalginross is accompanied by a c. 22 acre (c. 9 ha) temporary camp with Stracathro type gateways (St Joseph 1951, 64). Such camps have usually been assumed to be Flavian in the past (e.g. Maxwell 1981), mainly by association with nearby sites which are themselves often assigned to the Flavian Period by less than conclusive evidence. This dating may well prove to be correct, but given the evidence from Dalginross it might be safest if we were not to take it for granted until more positive dating evidence can be obtained from the camps themselves.

Finally, the fort also yielded ten fragments of Medieval pottery suggesting later occupation on or near to the site.

Background papers

Aerial photographs