It may be that it was due to changing military arrangements

that the closing years of the third century saw a renewal, after a

considerable interval, of pressure on the northern frontier, in the

shape of the Picts and Scotti. The threat appears to have brought

Constantius Chlorus (the western Caesar) to Britain twice – in

ad 296, when his principal concern was to return the British

provinces to the control of the central government, and again, 10

years later. The defeat of the British rebellion of Carausius and

Allectus is commemorated on the Arras medallion, which shows

Constantius disembarking at London and receiving the homage of

a personification of London; the legend proclaims him as ‘The

Restorer of Eternal Light’, presumably a reference to Constantius’

interest in the cult of the Sun. Constantius may also on that

occasion have set in train rebuilding work on the northern frontier.

It is evident that the rebuilding had to be radical in nature,

because of the decay that had taken place during the third

century. It is likely, too, that the rebuilding concentrated on the

forts rather than on the linear barrier and the smaller installations;

this appears to exemplify the new military imperative – defence in

depth, in which enemies might not be checked by a linear barrier,

but ultimately turned by a succession of forts which would be

encountered as they came further south. The rebuilt forts show

some changes from their pristine (Hadrianic) form: the regular

facing pairs of barracks were replaced by so-called ‘chalets’, many

of which were open-fronted, and with a variety of internal

arrangements. This may signify different manning arrangements;

possibly the barracks were for irregulars rather than auxiliary

units, and perhaps also they housed civilians from the vici as well. The logic of this presumably would have been that their security

in undefended settlements could no longer be maintained. There

is also evidence suggesting an ‘upgrading’ in size and standard of

the houses of fort-commanders, which may point to the beginnings

of a change in their status which became rather clearer in

the later years of the fourth century.

There is some evidence, too, that the defences of the Cumberland

coast were reactivated on a selective basis, presumably to deal

with a threat emanating from across the Irish Sea and to aid the

protection of the harbour of Moricambe (Kirkbride). It is noticeable

– though dating presents difficulties – that some blocking of gates

occurred at refurbished forts; this occasionally affected complete

gateways, but more often single carriageways. It is to be assumed

that the purpose was to make the forts less penetrable, so that, like

the forts of the Saxon Shore, they could serve as defended strongpoints.

Similarly, it has been suggested that, in some forts, former

internal towers were strengthened to support artillery-pieces.

The difficulties of the west and north evidently did not subside;

in c. ad 315, Constantine I assumed the titles, Britannicus Maximus,

which presumably point to success in a British enterprise,

although its precise application cannot now be recovered. During

the following decade or so, the west coast saw a programme of

construction of new forts, which followed a similar plan, although

less comprehensively applied, to those in the east, with high,

thick walls and bastions. Such forts are seen at Cardiff, Caernarfon,

Caer Gybi (Holyhead) and further north at Lancaster. It may also

be significant that at Lydney (on the Severn estuary) and at

Cockersands Moss (near Lancaster) we have evidence of the cult of

the Irish deity, Nodens (or Mars Nodens), suggesting that communications