As far as I know below is the most recent summery of the academic thinking and research of the history of the fort.
From Jill Eddison, Romney Marsh- Survival on a Frontier (Tempus Publishing 2000) pp. 48-52
"The fort we see today was first excavated by Charles Roach Smith and James Elliott (the engineer in charge of the Dymchurch Wall) in 1850, then by Sir Victor Horsley in 1893-94, and by Professor Barry Cunliffe in three seasons from 1976 to 1978. In 1981-82 Professor John Hutchinson led a geotechnical investigation into the effects of the landslips on the fort.
An area of some 4 acres (1.6ha) was enclosed by massive curtain walls with bastions at regular intervals. The walls were 12ft (3.65m) thick, with a core of Kentish Ragstone rubble, faced inside and outside with squared-off ashlar blocks of ragstone, with characteristic layers of red tiles laid at regular intervals between courses of ashlar. Most of the ragstone blocks were robbed away long ago, since they served as a ready-made quarry for medieval buildings, including Lympne Church and Castle on the crest of the hill above. The structure was however, wrecked by landslips and and has been further obscured by soil moving down the hill, so that although very substantial ruins remain, very little is in its original position (16). Three main landslips affected it. The most violent of these took place in the north-east corner, no doubt related to a large stream which comes down the hill. Another slip occurred in the south-east, dislocating the east gate and pushing a prominent lobe of the hillside out on to the marsh. The gate had originally been flanked by two bastions: the upper one had disintegrated, and the lower one had been tilted at a sharp angle inwards (17). The substantial base of the gate, built of large stone blocks, had been broken into thirteen fragments.
A third landslip was responsible for the collapse of the north walls. Trenches cut in 1981 slightly upslope from the remains of the north walls produced spectacular results, showing the Romans were well aware of the problems they faced in building on an unstable hillside. They had dug a foundation trench about 1.5m deep, and then driven closely-spaced short oak piles down into the Weald Clay – a not uncommon Roman practice when building on unstable ground. A platform of rough ragstone was then packed round the top of the poles to provide a working surface for the construction of the foundations of the wall. The story of the collapse of the wall unfolded in two of the trenches. Some of the oak piles had been bent when the wall collapsed. Those in one trench were now horizontal, and in the other they were also bent over at an angle (18). Above then, a trail of ragstone debris headed down the hill, above which was a shear-plane.
The wall had collapsed because the main landslip had removed the supporting ground from the lower side of it, while a mass of soil (colluvium) had built up against the upper side (19). It had slid downhill, bending the top of some of the piles as it went, and leaving a trail of debris behind it. Bastion 5 moved about 16ft (5m) downhill, while Bastion 6 moved only about 4ft (1.2m). Finally the wall tipped over, coming to rest at an angle between 22 degrees and 28 degrees. By establishing the original position of the northern walls Hutchinson was also able to prove the fort had originally been five-sided, with two upslope walls meeting at an obtuse angle at Bastion 6.
No trace of the south wall of the fort survives on the surface, but it was found when two trenches were dug through the lobe of the south-east landslip and out on to the Marsh. At the same time an important relationship was established between the relative timing of the landslips and the accumulation of sediment brought into the inlet by the sea. At the bottom of the trances a thick beach deposit of sand and shingle was found, and was clearly roman or post roman because it contained a fragment of red tile. Above that was 6ft (2m) of grey marine silt, and while that was accumulating, the clay ground on which the wall had been built slipped downhill and the wall toppled over. Three feet (1m) of the grey silt was deposited after the wall collapsed. It has not been possible to establish the date of the landslide that destroyed the fort, which is unfortunate because if that were known it would be an indication of the date of the silting of the inlet. But at least it is clear that the sea had accesses to the area for some considerable time afterwards. The fort itself was abandoned, for an unknown reason, in c350, some decades earlier than other Forts of the Saxon Shore."
