The steep rock of Mount Harşena makes a fortress easy to defend, but in the days of a siege what keeps a fortress alive is less the thickness of its walls than the water that never runs out inside them. The rock-cut water works of Amasya Castle, namely Cilanbolu, the Zindan cistern and the stepped passages said to link the fortress with the river and the royal tombs, are the most striking surviving answers to that need. The largest and most discussed of them is the Cilanbolu tunnel in the Upper Castle. For centuries it lived in popular talk as a well, a tunnel, a secret road and a depth without an end, and only the excavation and cleaning campaign of 2008 established its real dimensions and function.

The general history of the castle and the Ferhat Water Channel are treated in separate articles on this site. Here the focus is on the water routes inside the castle, and above all on Cilanbolu. Some figures that circulate in print for the tunnel differ from those in the excavation report, and the text below follows the report's measurements.

Location of the Castle and Its Water Structures

Amasya Castle stands on Mount Harşena, which rises north of the city centre on the northern bank of the Yeşilırmak. The rock of the castle lies about seven hundred metres above sea level and about three hundred metres above the river, and its walls, built of cut and rubble stone, descend to the riverbank in eight tiers. The history and sections of the castle, whose foundation is traced back to the Early Bronze Age, around 3200 BC, and which was repaired and kept in use each time it changed hands over the following centuries, are described in the article Amasya Castle (Harşena Castle).

The weak point of such a rock is water. Sinking a deep well on a mountaintop is impossible in most places, and going down to the river is dangerous under siege. The castle's rock-cut cisterns are the answer to this problem. Two structures are known by name today: Cilanbolu in the Upper Castle and the Zindan cistern on a steep slope northwest of the Kızlar Sarayı locality. Both were carved in the rock-tunnel technique, in a stepped form, and both have been cleaned in recent years by the Amasya Museum. Other galleries and water wells inside the castle are also reported. A stepped underground way running between the castle, the Yeşilırmak and the Royal Rock Tombs, thought to date from the 3rd century BC, is mentioned as well. Its present condition and exact course are not sufficiently documented, so part of what is told about it remains in the realm of tradition.

Official documents give the 1st century BC as the age of these structures. In the dossier prepared for Amasya's place on the UNESCO World Heritage Tentative List, the Cilanbolu cistern, the 186-metre Zindan Tunnel and the Ceylan Tunnel that reaches the river are described as monumental water structures with stone steps, cut as vaults into the bedrock by the rock-tunnel technique. The same dossier gives Cilanbolu a length of about 300 metres. That figure does not agree with the 205 metres of the excavation report, and the text follows the report. On the way up from the Royal Rock Tombs to the acropolis there is also a long, deep stepped tunnel said to have served the water supply; the path first continues northwest through a tunnel.

The reservoirs inside the castle were arranged for the days of a siege. The everyday water of the city below came from a separate structure, the Ferhat Water Channel. Dating to the Late Hellenistic or Early Roman period, this channel was built by cutting tunnels through rock, raising masonry in places and following the lie of the land on the principle of the water level. It is six kilometres long, and a stretch of about two kilometres running parallel to the highway at the Ferhatarası locality can easily be followed from outside. The story of the lovesick artisan Ferhat, who pierced the mountain for Şirin, belongs to a separate article and is not repeated here.

The Channels Described by Strabo

The oldest written information about Cilanbolu belongs to Strabo, the geographer from Amasya. Describing the fortress of his own city in Book XII of the Geography, he calls it a place so well devised that it can serve as city and fortress at once. In his account the rock is high and steep, drops abruptly to the river, and has on one side the wall enclosing the city on the riverbank and on the other the walls that climb to the summits. Two peaks, joined to each other by nature, are furnished with towers, and within this circuit stand the palaces and monuments of the kings. The neck connecting the two peaks is narrow, and the last ascent of one stadium from there to the summits is so sheer that no force can master it.

On water, Strabo's sentence runs as follows: inside the rock there are reservoirs, and the city cannot be deprived of this supply, because two tube-like channels have been hewn into the rock, one leading toward the river and the other toward the neck. The description fits the underground structures of Harşena so closely that nineteenth-century researchers relied on it when they spoke of the tunnel. In 1872 Georges Perrot, Edmond Guillaume and Jules Delbet wrote that the tunnel at Amasya was used to reach underground water, though they added that it was too large for that purpose. Which of today's structures correspond to the two channels Strabo mentions is not certain. Cilanbolu is associated with the description, but Strabo does not name the structure, and the excavation report treats his text only as the earliest written mention of the tunnel.

Strabo also shows the importance of a water store when he speaks of another fortress in the same region, Sagylion. On that high stronghold there was an ample reservoir that had been very useful to the kings. When Pompey ordered the garrisons pulled down and the strongholds made useless to brigands, the wells of the reservoir were choked with huge rocks. Arsaces, a son of King Pharnaces, fled up the mountain after attempting a revolt, found that he had no provisions and that the wells were blocked, and the stronghold fell not to force but to hunger. The passage shows that in antiquity the water stores of a stronghold were regarded as a direct target of defence and attack. The rock-sunk cisterns of Amasya should be understood by the same logic.

The Structure of the Cilanbolu Tunnel

Cilanbolu begins at a raised point in the middle of the castle, level with the great gate, and descends from south to north into the mountain. Its entrance is wide and round. The section that looks from outside like a majestic tower of cut stone leaning against the bedrock on its west side was reworked with rubble stone in later repairs. The structure starts in masonry and continues lower down by carving into the rock itself. According to the excavation report the slope of the tunnel is about thirty-five percent.

The figures that often circulate in print differ from one another. General accounts of the castle describe the tunnel as a shaft eight metres across entered by about one hundred and fifty steps, and some sources give its length as 181 metres, others as 186. The hundred and fifty steps describe the part that could be seen before the excavation. It is written that the steps were more numerous when the tunnel was built and that the lower ones have not survived, and local accounts too said the stairs disappeared further down, and indeed the lower part of the tunnel was blocked with collapse, earth and stone fill. The make-up of the fill was established in detail during the dig. Down to about the ninetieth metre the floor was covered by one to one and a half metres of earth, between the ninetieth and the hundred and tenth metres by medium-sized rubble, and between the hundred and tenth and the hundred and thirty-fifth metres by large rubble and, here and there, well-cut blocks. At the hundred and thirty-fifth metre, the last point reachable before work began, water could be seen among the stones.

Near the entrance, where the bedrock begins, there is an architectural fabric of the Hellenistic period, built of well-cut stone, round-arched and arranged like a gate. Higher up, continuing from it, one can follow a later repair whose walls are of rough-dressed and rubble stone and whose arch and vault are of brick and mortar. About thirty metres below the Hellenistic fabric there is a wall on each side of the bedrock, and a round arch of well-cut stone ties the two walls together; above the arch is an opening measuring 35 by 35 centimetres. The excavation team concluded that this masonry was built to support a rotten stretch of the rock.

In the middle section of the tunnel, near the ninetieth metre on the east side, stands a wall of bossed blocks, ten metres long and three metres high. Muzaffer Doğanbaş, who directed the work, interprets this Hellenistic wall as follows: when the tunnel was being built, a discharge shaft had been opened so that the debris cut from the rock could be thrown out easily, and when the work was finished the opening was closed with the wall. In some stretches of the tunnel, especially on the east side, water seeps out of the bedrock. Over time this seepage has produced karstic formations in one part of the rock. The water that collects at the bottom of the tunnel is evidently seepage from the cracks of the rock itself.

What Travellers and Local Historians Said

Cilanbolu also drew the attention of travellers who visited Amasya. Writers such as Evliya Çelebi, Charles Texier, Hamilton and Henry J. Van Lennep left information and comment on the nature of the tunnel. Among local sources, those who go into most detail are Mustafa Vazıh Efendi and Abdizade Hüseyin Hüsameddin, who carried his reports forward. In his Amasya Tarihi (History of Amasya), Hüseyin Hüsameddin relays a rumour current among the people: Cilanbolu was built as a water well for the castle, water flowing from Mount Karaman was channelled into it, this water was brought down as far as the river, and the place called Leğen Kaya was the end of this road. In his verse work on Amasya the same author mentions Cilanbolu as a well that arouses wonder in the middle of the castle, says that where it starts and where it ends are not known, and notes that its water is counted among the choice well waters of Amasya. In his account the well has no bottom.

Another report, which Hüseyin Hüsameddin relays from Mustafa Vazıh Efendi's work Belâbilü'r-Râsiyye fî Riyâzı Mesâili'l-Amâsiyye, concerns Caniklizade Mikdad Ahmed Pasha, the governor of Amasya (1776-1786), who hired a man to learn the depth of Cilanbolu. The pasha paid the man's legal blood-money in advance and had him lowered into the well, but the diver went into the water, vanished at the foot of the rock and never came out again. The story shows that for centuries the tunnel lived in the popular imagination as a deep and dangerous place. The 2008 excavation refuted both the claim that the tunnel descended as far as Leğen Kaya and the diver's story.

The 2008 Excavation and Cleaning Campaign

Rescue excavations in Harşena Castle began in 2006 and continued in 2007, and in 2008 the aim was to open the Cilanbolu tunnel. Under a permit from the General Directorate of Cultural Heritage and Museums, the work began on 2 June 2008 under the acting museum director Celal Özdemir, with a team of archaeologists and art historians including museum researcher Muzaffer Doğanbaş. According to press reports the Provincial Special Administration set aside 92 thousand lira for the job. The contractor first lit the dark tunnel, laid iron rails along the ninety-metre stretch filled with earth, and set up a conveyor belt outside the tunnel so that the material could be carried beyond the castle area. The fill was taken out in wagons.

Work began at the collapsed part of the tunnel, and at a depth of about one and a half metres steps of well-cut stone were reached. These stone steps, thirty-five to forty centimetres wide, had been seated in beds cut in the levelled bedrock and, in this part of the tunnel, were laid in rows of three. Some step stones were damaged and in places missing altogether. The section from where the dig began to the Hellenistic fabric was narrower and was arranged independently of the bedrock. After the Hellenistic arched fabric the tunnel widened and the stone steps continued according to the condition of the rock.

After the ninetieth metre the rubble fill began to be removed, and beneath it came earth fill again, with a layer of silt in places. At the same time the pool of water at the hundred and thirty-fifth metre was pumped out in two stages through plastic hoses. As the water receded, large cut blocks and rubble stones appeared at this point, and as they were taken away the work went deeper. The well-cut blocks removed from the tunnel were left outside it, to be used in a future restoration. At the hundred and seventy-fourth metre the stone-stepped floor was seen to end. Where the steps stop, a water reservoir was found, 6.90 metres long on a north-south axis and 2.75 metres deep. From the point where the work began to this spot a total of 356 steps were counted. Once the reservoir was found the excavation of this section was closed, and work moved to the southwest part of the tunnel to look for its starting point. There a collapsed stretch of 7.5 metres was cleared first, and the vault of the following stretch of about seventeen metres was found sound, with a height of 2.25 metres from floor to vault.

While the work was under way, press reports said that the clearance had reached a hundred and seventy metres and stressed that the tunnel had been opened in antiquity to meet the castle's water needs. Some local records write that the hundred and seventy metres were reached in 2009 and argue, from the castle's height of about three hundred metres above the Yeşilırmak, that it is not hard to guess the structure descends all the way to the river; the result of the excavation does not support this inference.The dig began on 2 June and ended on 10 October 2008, and it established that Cilanbolu is a tunnel 205 metres long in total. Potsherds of the Roman and Ottoman periods came out of the fill. The study finds include a Roman bronze coin, a Roman terracotta whistle and a fragment of a glass bracelet, two nickel Ottoman coins dated 1327, two further Ottoman coins, a terracotta pipe, a bone mouthpiece, a metal vessel and three pieces of a water jug. A metal key and an iron knife of the Republican period are also among the finds. The finds suggest that the tunnel was in use, or at least touched by human hands, from the Roman age to the twentieth century.

Conclusions of the Excavation

The verdict in Doğanbaş's report is plain: a considerable part of Cilanbolu was carved into the bedrock in the period of the Pontic Kingdom (301-47 BC), and its upper sections, built of rough-dressed stone, rubble and brick, form a large water cistern. The same report states that the claim of Vazıh Efendi and Hüseyin Hüsameddin, that the tunnel went down as far as Leğen Kaya, is not true, and that the diver story proved groundless. The tunnel has no exit at its end; at the bottom there is a water reservoir.

This finding also weakens a view repeated in some accounts. It used to be said that the first purpose of Cilanbolu was an escape route by which the king and his family could slip out of the castle during a siege. The dig produced no evidence that a tunnel ending in a reservoir and opening neither to the river nor to the outside was designed as an escape route. For that reason, regarding water as its main function is today the best supported view.

In some publications the tunnel is called "2300 years old". The figure corresponds to a rough age based on the Pontic period, and no exact year of construction has been documented. That the tunnel bears the traces of several phases has, however, been shown by the dig. The Hellenistic arched gate and the bossed wall represent the earliest phase, the vaults of brick and mortar the later repairs, and the Ottoman finds inside suggest that the tunnel supplied water until recent times.

The Stepped Tunnels of the Pontic Region

Cilanbolu is not an exception standing alone. In the Pontic region a total of thirty-nine stepped tunnels have been recorded at thirty separate places, more than one in some fortresses. According to the 2020 assessment by S. Yücel Şenyurt and Ahmet Emirhan Bulut, these are corridors cut into the bedrock that descend from the top of a fortress or natural rock height into the depths of the rock at a slope of twenty-five to forty-five degrees. They are found in central, northern and eastern Anatolia and are absent from the western coasts. Amasya Castle is reported to have three of them. The same study gives the length of Cilanbolu as 205 metres and counts it, together with the İncesu tunnel in Çorum (300 metres) and the Kurul Castle tunnel in Ordu (180 metres), among the large examples of the region. For the tunnel at Kurul Castle, about 1100 cubic metres of rock are calculated to have been cut away.

There are three views on what these tunnels were built for. The first and most natural is water supply. In the early 1900s Waldemar Belck suggested that the tunnel in Harşena might have been built to meet the castle's water needs and to store clean water. The second view stresses a military function. Hans Henning von der Osten argued that some tunnels in the bend of the Kızılırmak were too large for water alone and suited to descending unseen to valley level for a surprise attack; Guillaume de Jerphanion also regarded a military function as possible. The third view stresses a cultic function. Richard Leonhard said that so large a structure was unnecessary for water, that a small well would have sufficed, and he connected the tunnels with the cult of Cybele, the mother goddess of Anatolia. The small niches at the entrances of the Hisarkavak and Kurul Castle tunnels are cited in support of this reading.

Şenyurt and Bulut place the excavation results of Cilanbolu within this debate. Because it is firmly established that the tunnel has no exit and that a water-collecting reservoir lies at its end, in their view it is not possible to count the structure as a military part of the castle's defence system. Whatever the purpose, water lies at the centre of these tunnels. The reservoirs at the bottom appear designed to collect water seeping from cracks in the rock. In the cleaning at Tokat Castle, spring water met near the floor of the tunnel was not exhausted in about a month and a half even though one and a half tons were drained each day, which is cited as an example of how much water such reservoir systems can hold. The presence of more than one tunnel in the same fortress is explained either by insufficient water collecting in the first tunnel or by a population too large for a single tunnel to serve.

Craftsmen, Origin and Dating

Cutting rock on this scale demands a technique close to the knowledge of mining. The Pontic region is rich in iron, silver and copper deposits, and ancient sources speak of mining peoples called the Chalybes. Şenyurt and Bulut suggest that these craftsmen may have contributed knowledge and labour to the building of the tunnels. The idea that the five rock tombs of the Pontic kings on the slopes of Harşena and the stepped tunnels on the castle came from the hands of the same craftsmen has been voiced since the nineteenth century. According to Fleischer, the tombs are attributed in order to Mithradates I, Ariobarzanes, Mithradates II and III, and lastly Pharnakes I.

The criterion for dating rests on the period of use. Bronze coins of the time of Mithradates VI (120-63 BC) have been found in the tunnels of Geyraz and Kurul castles. At Cilanbolu the bossed wall at the ninetieth metre is dated to the Hellenistic age. From these data it is concluded that the construction of the tunnels cannot be later than the time of Mithradates VI. Whether an earlier origin exists is uncertain. Regional studies draw attention to the Hittite examples at Boğazköy and Oymaağaç (Nerik): underground structures built by corbelled masonry, without an exit and tied to spring water. Şenyurt and Bulut think the forerunners of the Pontic tunnels may be sought in this Hittite tradition, but a direct link has not been proved for Cilanbolu itself.

It is understood that some tunnels, after being used in the time of Mithradates, were filled in and closed by the Romans, but were later reopened because of their water sources and were used for water supply into the Ottoman age. The Ottoman coins, jug fragments and metal vessel at Cilanbolu agree with this view. At Tokat Castle, Evliya Çelebi's statement that the tunnel he calls the "Gazelle Road" descends to the river by 362 steps was confirmed when the cleaning begun in 2015 counted 361 steps in the tunnel. The comparison shows that stepped tunnels were still open in the seventeenth century and that Evliya really saw this structure.

Excavations Around the Castle and Its Present State

A year after the cleaning of Cilanbolu, in 2009, long-term scientific excavations were begun at Harşena Castle and Kızlar Sarayı under Emine Naza Dönmez, a member of the Art History Department of Istanbul University. Finds from Kızlar Sarayı belonging to a settlement under Persian rule and Phrygian cultural influence suggest that settlement around the rock-cut structures goes back earlier than the Pontic kingdom. The detailed results of these excavations are given in the article on the castle.

The most recent repair and restoration work on the castle was carried out between 2011 and 2013. Today Cilanbolu is a structure cleaned by the work of the Amasya Museum and opened to scientific research and visitors. A visitor looking from outside sees the tower-like masonry entrance of the first section, and those who go in descend by steps cut in the rock into a cool and damp darkness. In places in the tunnel, water seeping from the bedrock is a reminder that the rock of the castle still holds water today.

Chronology

Around 3200 BC: The Early Bronze Age, to which the foundation of the castle is dated.

301-47 BC: The period of the Pontic Kingdom, in which Doğanbaş says a considerable part of Cilanbolu was cut into the rock.

120-63 BC: The time of Mithradates VI. The stepped tunnels of the Pontic region cannot have been built later than this date.

Late 1st century BC to early 1st century AD: Strabo describes the water reservoirs inside the rock and the two channels.

1776-1786: The governorship of Mikdad Ahmed Pasha; the diver story is attached to this period.

1872: Perrot, Guillaume and Delbet write that the tunnel at Amasya was made to reach underground water but is too large for that purpose.

2 June - 10 October 2008: Excavation and cleaning of the Cilanbolu tunnel. The tunnel is found to be 205 metres long, with 356 steps and a water reservoir of 6.90 x 2.75 metres at the bottom.

2009: Systematic excavations begin at Harşena Castle and Kızlar Sarayı under Emine Naza Dönmez.

2011-2013: The most recent repair and restoration work at the castle.

2020: Şenyurt and Bulut publish their assessment of the origin and function of the stepped tunnels of the Pontic region.

Bibliography

Primary sources

Academic studies

Official sources

Other sources

Online sources