The field books of the Ordnance Survey were destroyed during the Second World War and it is not known why the location of the Aberffraw Palace shown on the O.S. map was so marked – see Fig.1.
In August 1957 Mr. Glanville Jones of Leeds University visited the site and reported that there was nothing above ground to suggest the existence of a building, but he agreed that available evidence was good enough to justify cutting some trial trenches.

The palace and buildings must have contained timber of sufficient size to be considered by Edward II to be worth moving to Caernarvon Castle in 1317. It is quite likely that Aberffraw Llys was almost entirely made from timber, in such a case, and in the light of its demolition, a building of this kind might well disappear without leaving any trace above ground. It was assumed that the timber uprights were either buried in the ground, and would leave post-holes, or alternatively they would spring from cill beams The cill beams may have been laid directly on the earth or on low stone sleeper walls.
Additionally, ploughing and general cultivation could easily have erased any evidence of the palace's existence. It is also possible that the site might have been partly covered with sand blown from the nearby dunes.

FIELD No. 914.

This was a small enclosure of half an acre then almost inaccessible from the remainder of the land constituting a small holding. In 1957 it presented a dismal picture; the waist-high grass and gorse bushes show that it was not agriculturally productive.
Trenches were dug and it would not have been unreasonable to have expected to encounter at least some artificial courtyard or road surface. There was not the slightest evidence of disturbance and not one piece of pottery or any other find was recorded.

FIELD No. 887.

The area around the site of the palace marked on the O.S. 25' map in field No. 887 was investigated  - see Fig 2.

A trench was cut through about 40cm of dark brown soil to the natural rock and a number of late sherds were found. It was clear that this field had been extensively cultivated in comparatively recent times. The broken pottery could have been driven out with dung, and pressed down deeply by cart wheels or hooves. A few metres away a terrace, common to fields 887 and 907, seemed a possible site for a building. Trench 2 revealed a natural rock platform with an approximately 10cm covering of turf.

Tench 3 was dug and  the rock was at least 60cm down. On extending this trench to the east a feature of some considerable interest was encountered. This was a rock-cut channel or water conduit about 20cm deep which was carefully covered with stone slabs which in turn were covered by a 5cm layer of clay, and a final covering of about  60cm of earth. This was clearly a very deliberate arrangement for conveying drinking water from one place to another and could in no conceivable way be associated with land drainage - see Fig 3.

The evidence then available certainly suggested that this is an early feature, and ipso facto connected with the palace site. Any late pottery sealed within the clay would certainly suggest a late date, but the purpose of the conduit would seem very difficult to explain. It is much more likely to be medieval and to have gone on functioning until long after the palace had been forgotten, which could explain the presence of late pipe stems in the channel itself.

CONCLUSIONS

Mr.Glanville Jones felt reasonably certain that field No. 914 did not contain the palace, and that further excavation would be unrewarding. However, he considered with the scanty investigation of the conduit, it would seem that the site of the palace must be very near.

In order of merit, fields 907, 887, 912 and 913 are candidates for the palace site. It is said that in 1926 two swords found in field 907 were presented to the National Museum. The report  stated that they are not in Cardiff, Chester or Bangor or the British Museum, and until they are produced nothing intelligent can be said about them.