Cisterns are installations for the storage of water and can either be underground or open reservoirs above the ground. Open cisterns (such as those of Aetius and Aspar) can be found in the belt between the Constantinian and Theodosian walls, since the land was devoted to agriculture. Water from open cisterns was therefore likely destined for supplying public fountains, industrial and agricultural purposes. Water for consumption was more likely stored underground.

Cisterns above the ground dealt with water pressure issues. Therefore, many ground-level cisterns were reinforced at the corners, which were typically angled or curved. Later cisterns also made use of convex ('curved' walls) reinforcement (Crow et al., 2008, p. 137).

Citizens were obliged to collect their water from designated reservoirs; diverting water from the aqueduct itself was severely punished (see Cod. Theod. 6.4.29).

By the Palaeologan period, many cisterns had lost their names (Crow et al., 2008, p. 22).

In the early period, water pressure of these open reservoirs was presumably controlled by sluice-gates. Crow et al., 2008, p. 131 note that the smaller closed ‘satellite’ cisterns (C3/2, D4/1 near the Aetius and Aspar reservoirs) may have been involved in regulating the flow from the reservoirs.

Open cisterns fell out of use in the Middle and Late Byzantine Period due to the destruction of 1204 and lack of maintenance (Altuğ, 2013, p. 120).

Water stored underground was better protected and more hygienic, and therefore more suitable for consumption. Drawing water was done sometimes in a similar way as drafting water from a well, but through a shaft.

They were predominantly brick-built (with some exceptions), and the interior was divided into bays. Supports were spaced generally 204 m apart; the roof supports were almost always columns. Building subterranean had another advantage: the surrounding earth contained the outward forced exerted by the water within. If a part of the cistern was above the ground, that part would need to be reinforced with extra buttresses. Sockets above the captials in the cisterns attest to the use of wooden tie-beams between the supports, possibly to ensure that the structure would move uniformly in an earthquake. Water in underground cisterns also needed to be areated. This could be done by inserting windows, which also provided a suitable overflow. The walls wee plastered to just above the height of the opening. Many subterranean cisterns possibly also served as the substructures of buildings (Crow et al., 2008, p. 137). The columns in underground cisterns could be recent manufactures, as well as reused spolia (mostly salvaged from damaged/destroyed buildings after i.e. fires and earthquakes). The majority of capitals (re)used date back to the 4th-6th centuries, when the Proconnesian marble quarries were in operation (Crow et al., 2008, p. 138).

Buondelmonti commented how many enclosed cisterns were ruined and had been converted for the growing of vines. Furthermore, by the Palaeologan period, many cisterns had lost their names (Crow et al., 2008, p. 22).

A consequence of out of order cisterns in decay was the accumulation of rubbish and sediment, reducing the original capacity of the citern, or hindering its entrance (Crow et al., 2008, p. 125).

'Castellum aquae' in Latin, 'taksim' in Turkish. Possibly 'kastellos' in Greek, see the description of restorations conducted by Romanos III (1030).

Allegedly according to Procopius' Secret History, people desperately flocked to the public fountains after Justinian willfully neglected the broken water supply of the city.

Fountain and water kiosk integrated in one building

'Su terazi' in Turkish. Siphon used to balance the pressure along the supply line. It worked like an inverted siphon: the pipe running up discharged into the water tank while the water decanted again into the conduit going down and this was repeated in sequences to the next tower until the water reached its destination (Kessener, 2017, p. 90, for further technical explanation, see Ibidem, p. 91-95). The water towers have been into well into the early 20th century, as Abdulhamid II completed the Hamediye water line. It is not known how long the suterazi kept functioning after this completion (Kessener, 2017, p. 106).

Byzantine period

Evidence for water towers in the Byzantine period is rather flimsly. A note in Crows work (2012) explores the possibility whether John Skylitzes' kastellos may refer to a water tower, a Byzantine equivalent of the later Ottoman suterazi. However, the meaning of kastellos remains too ambiguous, and according to Kessener, 2017, p. 86-87 it is more preferable to indentify these as castellum or public cistern. Only the remains of one possible water tower survive west of Antalya, however this specimen lacks a water tank on top of it. See for the discussion of "Byzantine" water towers Kessener, 2017, p. 87-88.

Ottoman period

It therefore remains unclear whether the suterazi was independently developed by the Ottomans, or whether they were preceded by a Byzantine version. They are however already depicted on Ottoman water maps (now in the Topkapı Museum), the oldest dating from 1584 and 1607. These water towers could either be repaired Byzantine ones, or newly built Ottoman ones. The Ottoman water infrastructure was mapped by Count Andreossy between 1812-1814 in his Book III of Constantinople et le Bosphore de Thrace (1828). It describes that the Ottoman water transport system consisted of piped conduits (lead or ceramic), running at or below ground level, and that at certain intervals it went up and down through a water tower with a water tank on top (Kessener, 2017, p. 89-90).