The Early
Castles of Gwynedd
Looking at the early Norman history of North Wales we have the fact
that Robert Rhuddlan in 1086 had been paying the king £40 per
year for North Wales, in the same manner as Rhys ap Tewdwr paid
£40 for South Wales. Although this has been argued
as ‘a speculative grant' by the king, there can really be no
doubt that such men would not pay such enormous sums per annum for
nothing but a vague promise. Certainly the historical and
archaeological evidence (what there is of it) supports Robert's
overlordship of North Wales from the late 1070s until his death in
1093. Similarly Rhys ap Tewdwr had begun his rule in
Deheubarth after 1078 and continued until his death some months before
Robert in 1093. Both men are liable to have been paying the
Crown their money rent for their Welsh fees during this period.
The archaeological demonstration for
Robert Rhuddlan being in North Wales comes from the evidence of castle
remains - in this case low-lying Norman motte and baileys. If we look for mottes commanding river crossings -
the traditional early Norman form of castle - we find
‘eleventh century' Norman castles that we have historical
evidence for at Rhuddlan, Degannwy (a reused Welsh
hill site commanding the Afon Conwy and not a real motte at all),
Aberlleiniog (a massive mound of uncertain provenance) and
Caernarfon (an alleged motte that Victorian excavation proved to be the
remains of a lime kiln turfed over and not a motte at all).
To this list can be added a second group for which we have no
historical evidence, but where physical remains and geography make a
Norman identification likely. These are Aber (possibly
referred to as Bangor in 1094), Nefyn, Dolbenmaen, Dinas Emrys
(again not a real motte), Dolwyddelan I and Pentre
Isaf, near Llangernyw. To this group can
be added Tomen y Mur which is a high lying site, in this case
being a reused Roman fort that carries a motte on its summit.
King Henry I (1100-35) was certainly here in 1114 and William II
(1087-1100) may have visited in 1098. In short, all of
Gwynedd and the western Perfeddwlad were encompassed by Norman mottes
which substantiate the historical evidence for Norman occupation as is
seen at the time of the Domesday Book.
If we look at castles that are likely to
be Welsh foundations we get some interesting comparisons.
There is evidence before 1200 of Welsh occupied ‘Norman'
fortresses at Rhuddlan (though quite where is another matter), Degannwy, Dinas Emrys and Caernarfon, while new Welsh fortresses had been built in
the west at Cymer, Cynfal, Deudraeth, Garn Fadryn and
possibly Tal y Cafn and Pentrefoelas.
Beyond this we are in the world of historical speculation.
What we can state with certainty is that Aber motte and bailey castle
shows no resemblance to the two Welsh built castles of Deudraeth and
Garn Fadryn. Both these are masonry structures lying
on rocky crags and ostensibly founded in the 1180s. The site
at Caernarfon is apparently gone, while both Rhuddlan
and Degannwy have no certain remains from this
period. It is a fair supposition that the site known as Rhuddlan Twthill never
received stone components, while even its existence as a castle is
suspect judging from the remains. Cymer was a small
motte on a promontory and was destroyed in 1116. The site has
a stone built eighteenth century house upon it. Whether this lies
on twelfth century castle foundations is impossible to say without
excavation. Cynfal was a motte surrounded by a rock cut ditch
and surmounted by a wooden tower that was burned down in
1147. Tal y Cafn and Pentrefoelas were both
apparently abarrant motte and bailey castles occupied and possibly
built by Dafydd ab Owain (d.1200/03) and destroyed in the late twelfth century by
his nephews.
The main sites without indications of
masonry defences in the supposed Welsh group of castles are
Aberlleiniog, Nefyn and Rhuddlan Twthill.
However, Aberlleiniog has a folly on its summit which may disguise or
obliterate any early structure, Nefyn is an alleged motte which has not
been excavated, but has been mutilated almost to the point of
extinction, while Rhuddlan Twthill has not been excavated and the sandy
mound looks an unlikely structure to have held up a wooden
keep. The pre-Edwardian castle of Caernarfon is unknown,
while Leland stated in the 1530s that the old castle had fallen into
the Seiont saltwater haven. The identification of its site as
being underneath the current Edwardian structure is therefore at best
debatable and based on no evidence, historical or archaeological -
especially when excavation and clearances make it reasonably certain
that Flint, Conway and Beaumaris were all built on virgin sites.
Of the masonry Norman castles Degannwy
is so ruined that nothing can usefully be said apart from the
round turret and wall to the north are reckoned, without evidence, to
be Welsh, but could just as easily be Norman. A round keep
and hall of the 1240s were excavated at the other end of the crag to
this. Dolbenmaen motte looks as if it once supported
a stone keep, while Aber castle motte most certainly did.
Both have fragments of wall core protruding from the motte
tops. Dolwyddelan I (Tomen castell) and Dinas Emrys both have traces of rectangular towers on their summits and
both most likely date to the tenure of Robert Rhuddlan and Earl Hugh of
Chester in the period before 1094 - although there is an unlikely
chance that both are Welsh built post 1100. The motte at
Tomen y Mur contains much good quality probably Roman
stonework and it is highly possible that this once consisted of a
Norman stone building that has collapsed. A similar
‘motte' made of a collapsed tower exists at the Welsh built
Prysor. The excavated rectangular tower at Dinas
Emrys was previously thought to have been a motte, which again
shows the dangers of judging a site without excavation. The
motte at Pentre Isaf is heavily overgrown although there are
some slight indications that stonework once crowned this feature.
The fragments of wall core projecting
from the periphery of the summit of Aber motte towards the south and
west, makes it all but certain that there was originally a small shell
keep or large round tower here. There is a tradition grown up
that Normans did not build round towers. This is simply a
theory and unproven in scientific terms. The Romans had round
towers as too did the Anglo Saxons and Normans. However, it
was certainly less usual for the Normans to build round towers, but
that does not exclude them from constructing the masonry on the mottes
at Aber and Dolbenmaen. Indeed it is far
from certain that such structures were round and not small polygonal
shell keeps of a common early variety. That the bailey of
Aber is virtually obliterated may suggest that it has been demolished
in antiquity. Certainly its position to the north is quite
demonstrable where no housing has been built upon the site.
To sum up, there seems little doubt from
the current evidence that Aber motte and bailey castle on the valley
bottom started life as a Norman motte and bailey castle built in the
decade before 1086. It was then destroyed during the uprising
of 1094 when all the castles of Gwynedd succumbed, the fall of some of
which are described in great detail in the History of Gruffydd ap
Cynan. The castle then seems to have lain
abandoned. It has then been claimed that after a hundred
years a mansion was built next to the motte and that this was the
palace of Llywelyn ab Iorwerth (d.1240). This claim has to be
examined against the evidence of what the excavations revealed and what
the documentary evidence actually states.
Garth Celyn
In November 1282 Prince Llywelyn ap Gruffydd dated a letter sent to the
archbishop of Canterbury at Garth Celyn. This place in Aber
parish has long been known to have been a major seat of the princes of
Gwynedd. Since the 1990s controversy has grown up over where that
site was. The evidence as suggested by site visits and research
of the original documentation - not mere antiquarian opinions - has
been summerised elsewhere. The modern condenter to be the site of
the llys of the princes of Wales is in the bailey of Aber castle.
Concerning this the following should be noted.
The site, locally known as the Mwd - and not Ty'n y Mwd - a name
fabricated from a nearby twentieth century house name and
disingenuously translated to the castle site for the 1993 excavation
report, GAT 1092 - appears a typical lowland Norman castle
site. The buildings excavated in its southern bailey show no
indication of a palatial residence, but of buildings that might be
found in a castle together with a post military industrial
complex. As a castle site Aber motte and bailey makes perfect
sense - a fortress to garrison 20 or 30 troops as well as followers to
command the crossing of the river and dominate Bangor - just 6 miles or
an hour away by horseback at a brisk pace. It is a further 8
miles from Aber to Caernarfon - 15 miles being a reasonable distance to
place between fortresses in a Norman zone of conquest - Degannwy being
only 8 miles from Aber as the crow flies, but including a dangerous
mountain pass and treacherous crossing of the Afon Conwy which could
add several miles. In short the earthworks and building
traces are all that might be expected for a castle. Could
then the ‘Norman' castle have been converted into the palace
or llys of the Welsh princes?
The modern idea that the buildings
excavated would have been a royal llys used by the Gwynedd royal
family, married 3 times into the family of the Plantagenets of England
and France, simply does not hold true. The small-scale
building in the bailey, with an industrial complex that cuts through
it[!], simply could not accommodate the royal family and court officers
that we know existed and attended the Welsh princes. There
would be the prince, his princess and the royal children, plus their
numerous body attendants. There would be 24 officers, 16 for
the prince and 8 for the princess. The most important of
these would be the captain of the household troops (who numbered 200
strong in 1258) and the royal priest - and we know that there was an
important free royal chapel at Garth Celyn llys because the king makes
note of it in the fourteenth century - stewards, falconers, justices,
grooms and the chamberlain. Plus all of these people would
have had their own attendants. When Llywelyn ap Gruffydd paid
a visit to the abbot of Basingwerk in the 1260s the abbot complained
that the prince came with over 200 people. This is when he
was on the move and not resident at his own primary llys when such a
figure could be expected to be higher with children and full time
servants as well as visiting uchelwyr. Aber castle bailey
simply does not have room for such an entourage. The building
in the bailey is claimed to be the royal Ty Hir mentioned in fourteenth
century documents. How can this be a long house when it is
not long by any stretch of the imagination? However the
masonry that can still be distinguished built into the house now known
as Pen y Bryn is obviously a long building and therefore could be
classified as a Ty Hir. The bailey site has revealed no cut
masonry - yet cut masonry has been found at Rhosyr and Pen y Bryn and
highly decorative freestone masonry has been uncovered at Degannwy,
Criccieth and Castell y Bere. All there is on the castle
bailey at Aber are river boulders laid in clay, with part of what
appears to be the fortress castle wall laid in a poor quality lime
binder.
Finally there is the word
Garth. This has been stated, solely when related to Garth
Celyn, to mean enclosure. No such usage of the word is
recorded in medieval Wales as a placename. However Garth is
commonly found throughout Wales and is universally applied to a jutting
spur coming out from a line of hills. Aber castle lies on no
Garth. The house of Pen y Bryn lies on just such a spur which
is marked on the oldest tythe maps as Garth Celyn - the projecting spur
of Celyn. In short all the evidence points to the enclosure
on the hill above Aber on the east side of the river as being Garth
Celyn and the motte and bailey in the valley on the west side of the
river as being the late eleventh century ‘Norman' castle,
whose bailey was later used for industrial purposes.
Copyright©2016
Paul Martin Remfry