Blake & Lloyd state that the Hwmyr is the boundary of Y Gogledd (‘The Left-hand’) in the Brut and that it is regarded as dividing Ynys Pridein into northern and southern portions. The find it curious that “we were unable to find any references in Welsh texts to the River Hwmyr outside of the Brut, and all references to the River Severn after 1136 refer to it as the Hafren.’ We have already seen how Hafren is the regular name for the Severn before 1136, with a definite occurrence of the equation Hafren=Sabrina in the Historia Brittonum of 829.
It is difficult, therefore, to regard the River
Severn as being referred to as Hwmyr at any time, no matter what the
influence of Geoffrey of Monmouth might have been on Welsh placenames. What
evidence do we have for the origin of Hwmyr, then? In British sources,
the only occurrence seems to be in the Historia Brittonum, where in
Chapter 61, we are told that ida filius eobba tenuit regiones in sinistrali
parte brittanniae, id est umbri maris ‘Ida son of Eobba held the regions
in the northern part of Britain, that is [north] of the Humber Sea’. There is
no evidence that the Historia Brittonum ever uses Britannia in the
restricted sense of ‘Wales’ that is occasionally found in Asser, a century
later, while the location of Ida’s kingdom is tied in to the next
‘confusion’, the location of Northumbria. Ida is regarded by all sources,
from Bede onwards (in other words, from 731) as the first English king in the
north and the founder of one of the kingdoms that was to become Northumbria. The
location of Northumbria is fixed not only by such things as placenames and
historical sources, but also by the occurrence at Jarrow of a building
inscription of Ecgfrith, king of Northumbria 671-685 and a descendant of Ida.
At the very end of his Historia Ecclesiastica (v.23), Bede summarises the ecclesiastical state of Britain in 731, listing the Bishops of the Cantwara, the East Saxons, the East Angles, the West Saxons, the Mercians, the people who live beyond the shores of the Sabrina, the Hwicce, the people of Lindsey, the Isle of Wight and the South Saxons. He then says that hae omnes prouinciae ceteraeque australes ad confinium usque hymbrae fluminis… regi aedilbaldo subiecti sunt ‘all these and the other southern provinces, up to the edge of the River Hymber… are subject to King Æthelbald’. There is no possibility of interpreting this passage as referring to the River Severn: Bede’s use of Hymbra clearly refers to what we now call the River Humber. It is also identical to the Historia Brittonum’s usage of Umber; the two words are linguistically identical. Hwmyr looks like a deliberate Gallicisation of Geoffrey’s Humber, an updating of the forms found in Bede and the Historia Brittonum.
Cair Efrog is one of the best attested names in early Welsh documents, just as Eburacum/Eboracum is in Roman Britain and the identity of the two names is usually taken for granted. However, they use a fifteenth-century poem by Guto’r Glyn that states yn Efrog yng Nglan Hafren ‘In Efrog in the Vale of Hafren. They recognise that, at this date, Hafren has to be the River Severn (although Blake and Lloyd believe that, before Geoffrey of Monmouth, it referred to the River Dee). The Brut names Caer Efrog as the burial place of Custeint, whom Blake and Lloyd identify with the Emperor Constantine, linking it with the modern placename Eaton Constantine in Shropshire, where they claim that there exists a tradition asserting that Constantine is buried in a mound near the village. As the village is close to the Romano-British city of Wroxeter, this enables them to identify the site with Caer Efrog.
It is difficult to get behind this farrago of implausibilites. The death of Custeint at Caer Efrog in the Brut derives from the Historia Regum Britannie (v.6), where Constantius dies at York; that this is not Geoffrey’s mistranslation is shown by Eutropius (Breviarium x.1), who states that Constantius… obiit in Britannia Eboraci ‘Constantius died in Britain, at York’. There is no mention of a tomb, though; this is found in the Historia Brittonum Chapter 25, where constantius constantini magni filius fuit et ibi moritur et sepulcrum illius monstratur iuxta urbem quae uocatur cair segeint ‘Constantius was the son of Constanine the Great and he died there [in Britain] and his tomb is shown by the town which is called Caer Saint (Caernarfon)’. The story of the tomb is simply that: a story. It has generally been assumed that it developed from a building inscription mentioning the emperor that was later mistaken for a tombstone.
So, we have no death of Constantine (who simply goes to Rome), an historical Constantius and a tomb at Cair Segeint. This makes a connection with Eaton Constantine in Shropshire something of a non sequitur. There are Roman temporary camps and an 8 ha fort at Eaton Constantine, probably dating from the early stages of the conquest of the West Midlands, but the name looks like a typical English placename with a manorial suffix (like Milton Keynes). The manor was held by Thomas de Costentin in 1242; this is a thirteenth-century spelling of Côtentin, in Normandy. The ‘tradition’ (for which Blake and Lloyd do not provide a reference) looks suspiciously like a folk etymology of very recent date to explain a placename whose true origin has been forgotten.
In short, the evidence for relocating Caer Efrog at Wroxeter is poor at best. In fact, Wroxeter has a perfectly good and well recorded Romano-British and Old Welsh placename: Viroconium Cornoviorum and Cair Guricon, both of which refer to the nearby prominent hill of the Wrekin. They choose not to mention this inconvenient fact.