Based on the archaeological excavations being conducted in Dunakeszi, we can create a more accurate image of how a fortlet -which can be classified as a late Román fortified port on the basis of its plan - might have looked like. The core area of suchlike symmetrical fortification was represented by a multistorey central tower, which was built on top of the bank, the highest spot of the surrounding.
The same scheme was applied in Dunakeszi, where the central tower was built on a natural relief of the riverbank. Two wing-walls ranging between 8 and 12 meters were attached to the opposite shorter walls of the tower, at the ends of which corner towers with thinner walls and square shape layouts less in extension, were constructed. From the angle towers two longer walls extended to the shore of the Danube. This type of forts can be familiar to us, since more of them were built in Rhineland and the southern portion of Dunántúl, in the quondam Valéria Province. Most of them have been recognized, however, in the Dunakanyar, e.g. at the present day settlements of Szob, Verőce and Tahitótfalu; their plans bear resemblance to that of the fortlet of Dunakeszi and by the fortifying ditch in the south and east, which has not been discovered so far. Nonetheless, we had the opportunity to unearth the southeastern corner tower that had been built on top of the high bank, which was characterized not by a developed a horizontal surface, onto which they developed a lower quality, 40 cm thinner sidewall for the tower made of disintegrative rocks being bound with weak mortar. A 20-cm-wide plinth protruded both on the internal and external sides of the tower from that ca. 2-m-wide wall extended to the central tower. The straight southern wall that ran from the corner tower toward the Danube had been planned to be 1.6 m thick. The upward remnants of it come at a 1.5 meters height at some points representing a spectacular and unique example among Roman ruins, the majority of which had been partially extracted down to the footing or perfectly quarried out. Upward remain of the northern regular square shape design, but a perpendicular trapezoid, 5.36 by 5.7 m in extension. The 1.3 m high sidewall was fairly meticulously strengthened up to 80 cm height from the protruding footing. By spreading the mortar evenly up to this height they wall presents another curiosity, namely that its width does not exceed 1.3 meters; thus it is 30 cm thinner than the foundation base and its southern counterpart, which ran in a 42 meters distance. The reason for this frugality, which was invisible from outside, might have been the economical use of building material, just like in case of the southeastern corner tower. A rampart platform from roof tiles was also built at the internal footing of the northern wall. It is a characteristic of the masonry technique that walls laid ir-regularly from stone blocks on both sides were supplemented with brick rows, and gaps were attentively filled with mortar. Due to the protruding basement footing that fitted perfectly to the slanting surface, we had the opportunity to reconstruct the contemporary landscape. Sides of the 1-m-deep foundation ditch, which had been dug into sandy subsoil, were strengthened with supportive piles. Imprints of these timbers are well visible in the plaster covering the foundation base. The enclosing 28-m-long remain of the southern wall is the longest of wall section ruins of the known fortified ports. The wall that has been reconstructed in full extension can be visited partly in the basement and in the backyard of the house at 28 Duna Row.
About the most important architectural unit of the fortlet, the central tower, very few information is known, as two residencies were built above it, almost completely overlapping its remains. Therefore the most essential source we possess about it is a brief report of the 1877 excavation of Flóris Rómer, in which he described the partially unearthed northern wall of the tower that lain in the still unbuilt area. On the external side of upward walls that extended hardly above the ground level he observed whitish plaster, the interstices of which were filled with dark red paint. During the 2009 excavation, suchlike plaster fragments were recovered in a great amount that had been painted red colour bands on in order to imitate ashlar veneer. Only the central tower had been decorated with architectural paint, since we did experience such traces on neither the southern nor the northern enclosing walls. Rómer discovered the floor of the tower, too, which were based on great size - presumably one and a half Roman foot large (45 x 30 x 7 cm in average) - bricks laid on sand that were coated with plastering mortar. It was a great surprise in 1985, when Roman walls with the height of a man were revealed in the cellar of 30 Duna Row, too. According to the location of the wall section, it could have only been associated with the north-eastern internal corner of the central tower, more accurately with its foundation. The foundation was entirely broken through when the former holder of the house drilled a passage for the cellar. Owing to this, we were able to measure the width of foundation structure that exceeded 3.5 meters. (Central tower of the fortlet at Dunafalva was characterized by the same width, whereas that of Verőce was only slightly thinner.) Upward wall of the tower might have been somewhat thinner, ca. 3 meters. The reason for constructing extremely thick walls lies not in the fear of an enemy's possible break through, since the neighbouring Sarmatians and Quadi were not aware of the siege technique that would have enabled them to occupy the fort. Building walls with extraordinary thickness provided op-portunity to employ higher supportive structures in order to increase defense. Based on our recent knowledge, external size of the central tower is estimated at 22 x 17,5 meters. We had a great opportunity to excavate the central tower to a limited degree in 2009, however, it did not yield any definite conclusion either. We discovered its southeastern corner in front of 29 Duna Row, on Római Street. We could identify a row of posts each measuring 10 cm in diameter, or more precisely, their imprints in the mortar fixing them. In the backyard of the plot at 30 Duna Row, we could only find the traces that used to have given place to the tower's disas-sembled northern wall earlier, which could have been at least 2.2 m thick at that point. The western wall measuring only 1.45 m was also found, whose foundation was constituted of large size rocks placed on one another in more rows without adhesive material. Conclusively, it appears to be probable that the construction works of the port was interrupted by the wartime incidents of 374-375 A.D., and accomplishment of the construction happened somewhat later than that, but thenceforth, with thinner walls of remarkably poor quality. It was also proved during the excavation of the tower that the central tower of the fortified port of Dunakeszi had been positioned asymmetrically in relation to the lines of the wing walls. It might have been a multistorey building that extended even a storey higher at the angle towers. Unfortunately we possess meager archaeological data to reconstruct images of forts facing towards the Danube. Since the river destroyed those walls that lain closer to the river, and thus were exposed more to erosion, as a conclusion of the movement of its bed.
As a result of deficiency in archaeological data, it was a common belief that forts were completely open between the lateral walls that ran up to the river, and ships moored at a wharf flanked by those walls. We already conceive it otherwise. A key to this issue might be the fort at Dunafalva that still had its riparian walls a century earlier. Nonetheless those walls have been completely washed away by floodings so far, we know it from past descriptions and drawings that towers were raised at and above each wall end at the riverbank, which were conjoined by walls that excluded the river. The towers that were indulged temporally were reinforced with compact wall up to the first floors at least in order that they were going to resist destructive effects of the river. Remains of the densely set supportive piles under the foundation of one of the lower towers were discovered in Verőce either. We should imagine the fortified port of Dunakeszi the same way. Gate of the fort might certainly taken place in the middle of the Danubian facade, which might have presumably looked on a wooden pier, in my opinion, as it is designated on the 3D reconstruction. A pole strongly shod at its end with iron, which we discovered on the bank in front of the fort, might have belonged to the structure of this wharf. The pier that was prevented by set stones from flowing direction had not been exposed to such a degree to the changes of water level and flooding as it is today due to regulations of the Danube. Paddle ships (lusoriae), on which a garrison fulfilled patrols and kept in touch with the opposite side, might have been moored to the wharf.
Finds discovered during the excavation bore witness to the ordinary days of those who lived in the fort. At the footing of the internal southern enclosing wall a thick layer of humus accumulated under the final rubble layer of the fortlet. The strata that had been aggregated since the completion of the fort consisted of abundant contemporary communal wastes, e. g. ceramic fragments, bones of consumed animals and/or others that were processed for industrial purposes, as well as Utilitiesand tools that did not function anymore. An excellent instance to this is an ornamented bone comb that was simply disposed of after it had been broken and thus became useless. The most comfortable way of wastage management was practiced inside the walls: they piled up garbage right at the walls or simply threw it over the parapet, Suchlike thick depositions of mingled humus and waste were accumulating in the forefront of the fort at Dunafalva during decades, which presented a real treasure mine of artefacts for archaeologists. Besides more than 80 coins, we recovered vessel fragments, unusable clothing accessories, iron nails and rich militaria finds from the excavation of small surface. As a point of interest, an abundance of female and children garments were recovered testifying that soldiers had moved to the fort together with their families. Suchlike explanation can be given to the existence of bone comb discovered at Dunakeszi.
The phenomenon is not unusual, since the presence of civilians can be detected in each late Román frontier forts. A tile was also recovered from the rubbish layer, on which a square grid had been scraped to serve as a board game. Onto an other brick an inscription was incised with erotic content. Both at Dunafalva and Dunakeszi, we collected numerous fragments of easily workable metals, dominantly lead, and other by-products that testified metallurgical activities in the fortlet. A lead plate found in Dunakeszi by Flóris Rómer did not belong to the roof structure but was raw material for manufacturing stud products. Traces of rough cleaving on its surface reflect to that type of procession. Another find, a recovered arrow head, can be an evidence for a raid against the fort. Charred wheat remains found in Verőce indicate that plenty of aliment was stored in the fortlets, which might have been commercial commodities likewise.
Based on the finds recovered at Dunakeszi and other defended harbours, we can reconstruct the later history of the counter fortifications along the Danube. It appears that destiny of the fortlets was not doomed by bloody sieges or barbarian destruction accompanied by fire, but it was a result of decay of the Roman Empire, and a gradual decline of its border defense. Leaving soldiers had not been ordered to set the forts on fire, or to destroy them, only to a complete abandonment. The peaceful abandonment is veri-fied by the cleanness of internal places and their deficiency in artefacts. Neither did Flóris Rómer find anything but stamped bricks and the lead plate. The inside of southeastern corner tower of the Dunakeszi fortlet was also perfectly 'swept-out,' all useful articles and equipments were taken away from it. We found no traces of boulversement caused by fights, or panic-like scuttle. Roof structure of the tower could certainly be remained complete even after its dereliction. From the massive layer of collapsed roof tiles found on the ground levelwe did not recover anything that reflected to intentional destroying (e.g. charred roof-beams or burnt tiles); therefore we may presume that the roof structure was collapsed after the wearing of timbers. Nonetheless, the tower was not completely empty in the period between the abandoning and the collapse of roof. Finds discovered lain on the ground surface reflect that some found shelter inside the empty building. A smaller fire burnt the floor right in the middle, which was surrounded by remains of a supper: animál bones thrown away and fragments of a broken large size storage vessél. Small speckle of the fire and the small number of finds indicate that the company could not have stayed in the tower for long, they left it soon.