Takamatsu Castle: A Fortress with a Sea Door

In an old photograph of Takamatsu Castle, water reaches the stonework below the Tsukimi turret. Its three tiers of roofs rise beside a low adjoining wall. Beyond the base lies open water rather than the reclaimed frontage that now separates the castle from the sea. The photograph, preserved by the Matsudaira family foundation and dated by the city to 1901 or earlier, records a relationship that the buildings alone can no longer fully show.^[1]

Nearby is Mizute-gomon, the water gate. A lord departing from here could enter a small boat, then transfer offshore to his larger vessel. This was an opening in the defenses whose purpose included taking the castle's master away.^[2]

Farther inside, water performed another task. It surrounded the main enclosure and left a bridge as its sole direct connection to the neighboring enclosure. Takamatsu placed restriction and departure within the same coastal complex. Understanding the castle requires following both: the water that interrupted a land approach and the sea route that began at a gate.

Ikoma Chikamasa began building Takamatsu Castle in 1588, after receiving the province of Sanuki on Shikoku. The site placed his seat beside the Seto Inland Sea. The later castle comprised successive enclosures and three moat circuits supplied with seawater. Its extent was greater than the grounds retained within today's Tamamo Park.^[3]

Enclosures divided the compound into manageable spaces. The innermost, the honmaru, contained the keep; the adjoining ninomaru formed another bounded area. Rather than surrounding one undivided block of buildings, the moats separated parts of the castle from one another as well as from the outside. A person already within the larger perimeter had not necessarily reached its most protected ground.

The sea added a different scale. A water obstacle might make crossing between two nearby banks difficult, while the connected maritime world made a distant shore reachable by boat. Takamatsu's arrangement exploited that difference. The castle needed places to stop a crossing and places to begin one.

These functions became especially legible under the Matsudaira family, who took possession in 1642. Their rebuilding changed the shore as well as the buildings upon it. The castle's coastal position was inherited; its later sea gate belonged to a landscape the new household enlarged.

From 1671, land was reclaimed to form the northern enclosure, the kitanomaru. The Matsudaira were extending the castle over ground that had previously been sea. A joint in the masonry still identifies an addition beneath the Watari turret: different stretches of stonework meet where the enclosure grew.^[4]

The extension changed which line counted as the edge. A sea gate serving an earlier boundary could become redundant when new ground appeared beyond it. At Takamatsu, part of that earlier arrangement was absorbed into the later building. Three southern bays of the Watari turret reused members from the old Umite gate, and investigations found former gate foundations beneath the older stretch of wall.^[2][4]

This was an adjustment of a working perimeter. An older opening did not have to remain an opening for its material to remain useful. The enlarged enclosure supplied another sea-facing entrance, with a turret, connecting buildings and a landing relationship appropriate to its new position.

The Tsukimi turret was raised in 1676. Mizute-gomon stands beside this northern group, though investigation indicates that the gate now surviving was probably rebuilt late in the Edo period. The sea approach therefore continued through changes to its fabric. It cannot be dated as one untouched assembly constructed by the castle's sixteenth-century founder.^[2]

On the seaward side, excavation recovered evidence of a stepped facility. The stone steps presented there now are reconstructed. Their significance is the transition they mark: from firm ground within the castle to a level where someone could embark. A gate could regulate passage through the wall; steps had to make the next stage of the journey physically possible.^[4]

The old photograph restores the scale of that transition. Water spreads away from the masonry, and the turret stands directly beside it. Reclamation had created this castle edge in the seventeenth century. Further reclamation outside it would eventually alter the edge again.

The lord's larger vessel did not need to fit against the water gate. The documented arrangement used a small boat for the first stage, followed by transfer to the gozabune, the vessel fitted for the lord, offshore. That division connected a relatively confined embarkation point to a ship with space for a much more elaborate household setting.^[2]

Takamatsu's best-documented example is Hiryū-maru. Matsudaira Yorishige ordered the first vessel of that name in 1669. It was subsequently replaced several times. A surviving set of eight drawings concerns two later rebuildings: three sheets from 1754 and five from 1792. They preserve different versions of the named ship rather than the unaltered appearance of the first hull.^[5]

One 1754 sheet presents the whole vessel. A large sail rises above the hull; red and white hangings define the occupied space beneath it. Crests and the ship's name make its affiliation visible, while weapons appear alongside the decorative display. The composition brings the vessel's parts together for inspection. It need not be read as a record of one actual departure in a particular wind.^[6]

Other sheets enter the accommodation. Painted partitions distinguish rooms within the raised structure, giving the vessel an interior vocabulary recognizable from a residence ashore. A plan associated with the 1792 rebuilding identifies a steam bath and a lavatory. Traveling involved bodily needs as well as a public appearance; the drawings assign those needs places within the ship.^[6]

The comparison with the water gate is revealing. The gate and small boat managed the transfer from shore. The larger vessel supplied a setting in which the lord could remain while moving over a greater distance. Embarkation did not mean stepping straight from a palace room into another room of equivalent scale. It meant passing through different spaces and different craft, each answering a different part of the journey.

Such journeys included alternate attendance, the obligation to spend recurring periods at the shogun's court in Edo. For a lord based on Shikoku, leaving the domain involved crossing water. The castle's sea door connected his residence to a route required by his political position, as well as to excursions closer to home.^[5]

That access also depended on arrangements beyond the gate. The domain maintained a harbor and places to keep its vessels on the castle's western side. The Kagawa Museum traces an enlargement of the dedicated harbor between 1656 and 1670.^[5] The gate was an embarkation point within this larger maritime establishment, not the berth and storehouse for everything that sailed from Takamatsu.

The route into the main enclosure posed the opposite problem. Here the moat enclosed a relatively small piece of ground, and Saya Bridge was its only direct link to the ninomaru. The keep rose within an area reached through a limited crossing rather than from a broad forecourt.^[7]

A castle plan of 1627 already shows a bridge in this relationship. Early representations describe or depict a bridge with railings. By an 1823 repair drawing, the passage had a roof. The municipal bridge-repair report places these images in sequence without supplying an exact year when the first roof was added.^[8]

The covered passage gives the surviving arrangement a particularly enclosed appearance, but the older relationship did not depend on a roof. Water confined entry to a particular crossing. A wall and gate at its end could then regulate movement onto the next piece of ground. The physical sequence made access divisible into stages.

It is important not to insert every bend of the modern visitor route into that earlier sequence. The city's comparison with an old plan shows that the Edo-period approach beyond the bridge differed from the present longer detour. The bridge itself also underwent movement and substantial repair in modern times.^[7][8] The enclosing water and the restricted connection are clear; a detailed reconstruction of an attack cannot be read directly from today's path.

Seen together, bridge and sea gate explain two different kinds of control. At the honmaru, a bridge concentrated an approach from an adjacent enclosure. At Mizute-gomon, an authorized departure opened onto a much wider range of possible destinations. The walls did not turn all movement into the same problem. They separated an internal crossing from access to the sea.

The keep itself was dismantled in 1884 because of deterioration. Its stone base preserves the location of the tower without supplying its missing upper structure.^[3] Even without that silhouette, the relationship between inner ground, surrounding water and the crossing remains available to understand.

Some of Takamatsu's relationship with the sea was carried in trained bodies rather than boats. The city's account of the swimming tradition later called Suinin-ryū dates organized instruction for retainers to 1643. Yorishige directed Imaizumi Hachirōzaemon to teach swimming at the waterfront training ground west of the castle. Summer demonstrations were watched by the lord.^[9]

The castle's own inner-moat history also records that Yorishige swam there in the year he arrived. The wider training ground and this episode within the moat were different locations, but both placed entry into water among the activities of the ruling household.^[10]

These records add a useful qualification to the language of defensive barriers. A moat obstructed ordinary passage on foot; it was not a claim that water made every human crossing impossible. A coastal retainer could be expected to acquire skills for conditions in which a bridge or dry path was unavailable. The same environment could demand a barrier in one setting and competence in crossing it in another.

The surviving swimming tradition does not establish that a particular maneuver was used during a siege of this castle. Its more direct contribution is to the history of preparation and display. The lord's maritime setting involved ships and embarkation structures, but also repeated instruction in what a person could do once in the water. That is a different kind of adaptation from adding another wall.

The surviving moats retain a connection with the sea. In the arrangement described by the city, water passes through the sluice between the ninomaru and the sannomaru, or third enclosure. Marine life in the moat makes that connection conspicuous. Seawater belongs to the functioning site, not only to its historical name.^[10]

Yet the modern sluice creates a trap for historical explanation. It would be easy to describe the visible mechanism and assume that the seventeenth-century castle operated it in exactly the same position. The city's investigation gives reasons not to do so. An old comprehensive castle drawing shows a defensive gateway arrangement rather than the present sluice, and the adjacent masonry has evidence of alteration. The historical method of managing the moat water remains unresolved in that account.^[10]

Two parts of the interpretation can therefore proceed at different levels of precision. The castle's use of a coastal, seawater-filled moat system is well established. The location and operation of its earlier water-control connections require more evidence. Knowing the former does not supply the missing details of the latter.

For the sea gate, excavation and the surviving shore-facing buildings identify an embarkation relationship. For the sluice, a functioning modern connection does not by itself recover an Edo-period installation. The difference keeps the castle's practical history open to investigation without reducing the whole maritime setting to uncertainty.

It also helps explain the changed frontage. Land reclamation separated the old north wall from open sea, while a water connection was retained. The coastline could move outward without turning the remaining moat into an ordinary landlocked pond. The form of the connection changed more than the substance flowing through it.

The household ashore changed its position too. In 1700, the domain's palace moved into the sannomaru, an enclosure previously used in part for warehouses. Residence and government were consequently not confined to the island-like honmaru or to the keep. The larger castle provided other ground on which the household could expand its daily business.^[11]

The Hiunkaku standing in that enclosure belongs to a later period. It was built between 1914 and 1917 as a villa for Count Matsudaira Yorinaga, a descendant of the former ruling house. The earlier domain palace had also been called Hiunkaku, but the shared name does not make the present building an intact residence from the age of Hiryū-maru's eighteenth-century drawings.^[11]

Other retained buildings underwent different changes. The Ushitora turret was built in 1677 at the northeastern corner of the eastern enclosure and moved in 1967 to the former Taiko turret site. Its old structure survives in a new position. Sakura Gate, at the entrance to the sannomaru, burned in 1945; its replacement was completed in 2022.^[2][11]

These alterations affect how the remaining ensemble can be read. A turret may retain old fabric without identifying the original edge of the castle. A new gate can restore the appearance of a lost entrance. A villa can preserve the family's later presence without continuing its former government. Takamatsu became a public city park in 1955, after the city acquired it the previous year.^[3] The places through which visitors move are the result of these different decisions, not a single surviving household left ready for departure.

The old photograph holds the Tsukimi turret against water extending beyond its base. Its principal historical detail is the edge itself. It records where masonry ended and the sea began, before a larger reclaimed frontage intervened.

The ship drawings supply what the photograph lacks: the accommodation that waited beyond an embarkation point. The bridge plans supply another relationship, concentrating entry to the main enclosure. Together, these materials distinguish a sea-facing threshold from an internal crossing. Both belonged to the castle, but they did not serve interchangeable journeys.

Takamatsu's defenses took advantage of water without making separation their only purpose. A lord could be protected within successive enclosures, cross to the northern gate, and leave by boat for obligations elsewhere. His residence was connected to an offshore vessel and a maritime route through an ordered transfer, rather than through an uninterrupted extension of dry ground.

That route no longer operates from the gate. What survives is uneven but specific: the sea-facing entrance, the turret beside it, archaeological evidence for steps, the isolated keep ground, and moats still connected to seawater through an altered arrangement. The former shoreline can be recovered in a photograph; the former ship in drawings. The castle becomes intelligible when the different crossings are put back in relation—not when every remaining stretch of water is assigned the same job.