Tretower castle may have been commenced before 1081 when King Caradog ap Gruffydd, the prince of Ystrad Yw, and his Norman followers were overwhelmed by Rhys ap Tewdwr and Gruffydd ap Cynan at the battle of Mynydd Carn.  It was almost certainly in existence by 1103 and held by Picard, the first baron of that name.  Before 29 October 1233, Earl Richard Marshall of Pembroke and Chepstow had taken Roger Picard's castle of Tretower and thrown it to the ground.  As a consequence of this attack it is suggested that Roger commenced a refortification programme which included the building of the great round tower within the recently ruined shell keep.  Whether this reconstruction included the building of the bailey and its gatehouses is unknown.  Sometime, probably shortly after 1234, Roger made a grant to St John's priory, Brecon, of a parcel of land south of the town square beyond the east gate of his castle which was then known as the Boket, or west gate.  This shows that 2 bailey and gatehouse existed by 1234, but proves nothing else.

During the Welsh War of 1244 to 1246, the castle held out when Brycheiniog was attacked, but,
in December 1262, the district then held by Roger the son of the above Roger Picard, swore homage to Prince Llywelyn ap Gruffydd (d.1282).  Despite this act of treason to King Henry III and Roger's subsequent support of the Reformists under Earl Simon Montfort of Leicester (d.1265), there is no mention of the castle changing hands and it was definitely still held by Roger in 1273 even though he had 'adhered to Llywelyn'.

In 1308 the family of Picard of Tretower died out and the castle and lordship passed to the Bluets of Raglan, who seem to have begun Tretower Court as a manor house.  In September 1403, their descendant, James Berkeley, was ordered to munition the ruined castle and this he seems to have done by building a wooden storehouse between the ruined walls of the great tower and the shell keep.  With the repression of the revolt in the Marches the castle was no doubt allowed to continue its long process of decay and when Leland passed by in the 1530s he mentioned the court, but not the castle.

The castle keep is said to stand on a low motte, which externally is only about 3' high from the surrounding ground.  This has a base diameter of about 100'.  There is a shallow, 30' wide, ditch surrounding the mound, the making of the one, allowing the creation of the other.  When the castle was built here, in low-lying, waterlogged ground, the builder ignored the previous Roman fortification at nearby Pen y Gaer and the civil and religious settlement at Cwmdu.  The ditch would appear to have been filled with water from the surrounding brooks that fed into the nearby Afon Rhiangol.  It is possible that there was no wooden phase at Tretower and the mound was instantly reveted.  Some apprently early reveting can still be seen underneath and in places encased by the later masonry of the shell keep.  Presumably a bailey was constructed at the same time, probably with wooden defences, though this cannot be proved.  The ward lies to the east and is about 200' N-S by 160' long.

It was probably during the Anarchy (1135-54), that Roger Picard succeeded his father and also acquired the nearby lordship of Llansantffraed which was later to be known as Scethrog.  It has been suggested that Roger may have used his enhanced wealth to build the ornate shell keep in his massively expanded lordship, but of course this is pure guesswork.  The shell keep consisted of an external wall up to 30' high with a square gatehouse to the east.  This, roughly 20' square tower, was slightly smaller than the similar one at Hay on Wye castle.  The shell keep used the thin revetment of the earlier motte as its batter and had thin walls just under 6' thick.  Such thinness is probably indicative of an early date when siege weapons were not so powerful.  As such there is no evidence to say that the keep did not date to the earliest years of the Picard conquest of the district.

The new shell keep seems to have initially consisted of a residential block running E-W on the south side of the motte.  This side also had a five sided projection which seems to have always housed the castle kitchen.  The fireplace in the kitchen with its ingenious smoke dispersal system is well worth examining.  On the west wall of the shell was built a long solar range, the 'motte' being lengthened and straightened to carry it.  From the north end of the solar range the shell wall ran in five uneven segments to the once tall squarish gatetower that in the days of Buck still dominated the castle site.  This lay next to the hall block.  Buck's print of 1741 shows the shell still standing to its full height, with the battlements apparently rising to the taller north solar side of the keep.  All of this wall and the square gatetower have now collapsed.  A list of other shell keeps appears in the discussion on Tonbridge castle.

After the castle's destruction in October 1233, the shell keep seems to have been gutted and the new round tower, perhaps based on that at Bronllys, begun.  The original tower seems to have been intended to have consisted of a basement with two storeys above.  This was later altered and a third floor was added.  Externally the round tower had a sloping plinth and a moulded string course girding its 9' thick walls.  Entrance was initially gained to the round tower via a first floor doorway reached by a wooden forebuilding of which the steeply pitched roof crease can still be seen.  A similar arrangement was used at Bronllys, Skenfrith in Gwent as well as Longtown in Herefordshire.  It should be noted that there was a chamber in the wall of the shell keep opposite the entrance to the round tower and that this may have been part of access to the inner keep.  Another tower keep within a shell keep can be seen at Launceston in Cornwall.  In the thickness of the wall between two doors of the inner keep is a mural stairway which leads sharply upwards to a spiral stair which gave access to the upper floors.  In the window opposite the entrance is a mural stair that curves down allowing access to the basement.  The holes for rafters are very apparent between the various tower levels and on the second storey to the north, is a doorway that gave access to the shell keep wallwalk, probably via a wooden bridge or passageway.  Similar arrangements can be seen at Chateau Renault and Freteval castles west of Paris, where again a round tower within a shell keep.  The roof and rafters associated with this door can be seen to date from a different period in construction as they overlay and cut into the relieving arch and cut stone of the doorway.  Probably this was the wooden structure built by James Berkeley in his munitioning of the castle in the early fifteenth century.

Work on the third floor of the round tower seems to have come to an abrupt halt as is seen by the offset just above shell keep level.  This offset has cut through a loop in the spiral stair to the east.  The lower part now makes a fine rectangular loop and is topped by the coursing layer.  Above it the opening is made into a round headed window.  Obviously building work stopped at this time and was then restarted at a later date.  Perhaps a gift of wood to Roger Picard in 1245 marked the completion of this tower after a ten year hiatus in its building - however wood was used for many purposes in that era.  Judging from Buck's print yet another storey stood upon this tower before the battlements were reached.  No trace of this storey or the battlements now remain.
 A list of other round tower keeps can be found in the description of Llanstephan castle and a general list of round keeps is found under Dundrum. 

To the east of the shell keep lies the crumbling remnant of the bailey defences, now occupied by a farmyard.  The remains of this and Buck's print suggests that a large drum tower stood at the two extremes of the triangular ward, that to the north still being standing in 1741.  The buttresses supporting the south wall still largely survive today, but the apparent postern near to the keep does not.  The entrance between two buttresses set between this postern and the south tower was probably relatively recent in 1741 and has now been blocked again, the wall here still standing some 6' high.  Other parts of the enceinte stand 20' high.  Presumably the Boket gate stood to the SW as it was said in the c.1234 charter to lie to the west.  Buck's print also seems to show evidence of what may have been a mantlet wall surrounding the keep on the other side of the moat.

Why not join me at Tretower and other British castles this October?  Please see the information on tours at Scholarly Sojourns.


Copyright©2016 Paul Martin Remfry