Hakata Bay's Stone Defence Line

Even before 1274, the shogunate had ordered warriors with holdings in Kyushu to undertake regional defence. Northern Kyushu was therefore not an unguarded blank. What changed after the invasion was the addition of a long physical structure to existing and reorganised forms of coastal guard duty.

Fukuoka City's survey history sets out a sequence: foreign-defence guard duty was organised in 1275, followed by construction of the ishitsuiji in 1276. This chronology resists the misleading image of a stone wall appearing immediately and independently after battle. Guard organisation and construction were connected preparations for another invasion.

The conservation plan dates the main work approximately from March to August 1276, a period of about six months. That duration and the roughly 20-kilometre line indicate the scale of the undertaking. They do not, however, provide enough evidence to calculate a daily construction rate or a precise total workforce.

The barrier was not drawn across level urban land. It followed dunes and coastal slopes, responding to sand, nearby stone sources, and local topography. Those variables are essential to understanding why excavation does not reveal one repeated cross-section along the entire front.

Different categories of evidence illuminate different stages. Orders can show that construction or repair was required; excavated masonry can show materials and form; earthworks and ditches reveal supporting elements; pictorial evidence can place people near the barrier. No single category provides the full line by itself.

The image of labour in this scene is an AI-generated reconstruction. It symbolises moving and stacking stone on dunes, but it does not establish the number, clothing, tools, schedule, or identities of the workers. The documented period and landscape are retained while unknown labour details remain unknown.

At Imazu, excavation found a trapezoidal stone structure on the seaward slope of a dune, approximately three metres high, two metres wide at the top, and three metres wide at the base. Sand filled internal gaps without clay. Granite predominated in the west and basalt in the east, with alternating material in a central zone.

At Ikinomatsubara, a stone structure descending towards the sea measured roughly one to one and a half metres wide and survived to about 1.8 metres high. Clay reinforced its landward side. Pegmatite-bearing granite was common in the west and sandstone in the east; Fukuoka City notes that the material change may relate to the boundary between Hizen and Higo responsibilities.

At Nishijin, one excavated example used a clay bed to stabilise the dune before masonry with a base width of 3.4 metres was built. Another investigation found an earthwork behind the stones with a gap of about one metre, creating a double structure. The Imazu section cannot simply be copied onto Nishijin as a universal design.

Hakozaki differs again. Alongside its masonry, a large ditch ran on the landward side. The conservation plan treats stonework, earthworks, and ditch-like remains as elements expressing the site's fundamental defensive value. Comparing only wall height would miss these related structures.

Variation does not automatically mean poor workmanship or failed standardisation. It can be understood as adaptation to terrain, nearby material, assigned groups, and reinforcement needs while pursuing a shared aim: creating an obstacle along the coast. Unknown or unexcavated sections cannot be forced into the four known patterns.

The scene image visualises different stones and cross-sections but is not an archaeological drawing. Dimensions belong to the text and cited reports. The number, colour, and placement of individual stones in the AI-generated visual must not be read as measured archaeological data.

Five years after construction began, Yuan forces again approached northern Kyushu in 1281. Unlike the coast encountered in 1274, Hakata Bay now had the ishitsuiji and an organised guard system. Fukuoka City's plan and museum explanation assess that the barrier helped prevent a landing along the bay.

Prevention at the shore does not mean that one wall ended the entire campaign. The barrier changed the conditions under which troops could establish themselves after reaching land and worked in relation to defenders. Its role should be defined at that point of contact rather than expanded into a total explanation of the war.

The illustrated handscroll commissioned by Takezaki Suenaga includes a scene in which Suenaga and his followers pass in front of the barrier at Ikinomatsubara. This is important pictorial evidence that people and the structure were joined in a near-contemporary representation. One scene cannot reveal every section or action around the bay.

Weather inflicted major damage on the invading fleets, while the barrier also performed a role in obstructing a landing. These propositions need not compete as mutually exclusive explanations. They refer to different events, places, and moments within a complex campaign that also involved fighting and movement at sea.

If the outcome is attributed only to a storm, the five years of guard duty, construction, assigned responsibility, and maintenance disappear. If it is attributed only to the barrier, maritime action and weather disappear. The evidence is better served by stating the supported function of the barrier without converting it into a single cause.

The night scene is an AI-generated reconstruction of guards on the coast and ships offshore. It does not reconstruct a particular date, number of vessels, order of attack, named warrior, or weather condition. It visualises the relationship between guarded coast and seaborne threat.

A 1316 implementation order issued by Shoni Sadatsune states that the ishitsuiji at Hakata's foreshore was damaged and required repair. Roughly four decades after its construction, it was still treated as a defensive facility that should be maintained. A narrative confined to the single campaign of 1281 would miss that duration.

The conservation plan finds evidence of maintenance connected to foreign-defence guard duty into the first half of the fourteenth century. Repair can be read as part of an enduring system rather than merely replacing fallen stones. The evidence does not show that every section was repaired at the same interval or in the same manner.

As the external threat receded, the barrier ceased to function as an active defence. Sections were buried in dunes, stones were removed and reused, and development made much of its course difficult to recognise. The absence of a continuous visible wall today is a result of long change, not proof that a wider line never existed.

Fukuoka City Museum identifies an excavation at Imazu Nagahama in 1913 as the beginning of modern archaeological investigation. The site received national historic designation in 1931. From 1968 to 1971, major investigation and conservation work proceeded at Ikinomatsubara, Imazu, Nishijin, and other locations.

What visitors encounter can include exposed remains, restored masonry, protective covering, and explanatory facilities. A reconstructed section should not be treated as untouched medieval fabric, while a protected or buried section should not be treated as empty ground. Seeing the site requires attention to both archaeology and conservation.

The true length of the Genko Borui cannot be understood from a line of stones alone. The 1274 landing, provincial assignments, local building methods, patrols, the 1281 defence, the 1316 repair order, modern excavation, and preservation form a longer line of institution and memory across the same coast.