The dimensions of the cisterns are extremely variable and depend directly on the function they served. They can range from a few cubic metres for a private dwelling to several thousand litres for a public building, such as the Graeco-Roman baths of Kom el Dikka. (Graeco-Roman era).

The single chamber, small volume reservoirs certainly correspond to domestic cisterns storing rain water or perhaps filled, as Gratien-le-Père remarks, "by water skins carried by camels, mules or donkeys".

This latter method was also employed to provide water to the Ottoman town, which was not equipped with a dynamic network. Likewise the description of the method of filling the cisterns of Fort Qaitbay, as noted in 1611 by George Sandys of the English Foreign Office :

"On the isle of Pharos, now part of the mainland, there stands a fort that defends the entry to the port; there is no water except for that which is brought by camel from the cisterns of the town…"

The large volume cisterns must have been destined for public use. The medium sized could have been to supply a district or group of houses.

Type V, built on several levels, and much in evidence in the Kamil files, seems to belong most specifically to the Arab period of the town. As for those cisterns situated the deepest in the ground, they would correspond to the earliest levels of occupation of the town; that is, the Graeco-Roman era.

In general, and for those that can be studied, the cisterns hardly show any trace of the system of supply. However, the texts are clear and they recount more or less the same thing. The cisterns were filled at the time of the Nile flood, between August and September, with water brought from the canal through aqueducts and secondary channelling. As Father Vansleb, who visited Alexandria in 1672, remarks, the connection with the cisterns is done by an "ingenious method". But what was this method?

As evidence of the reality of this supply system, we have only one plan drawn up by a Greek Alexandrian architect, Paraskevas. This precious document dates from 1927 and plots, in 1/100 scale, the cisterns and underground tunnels encountered during the building of foundations of the Alexandria Water Company premises.

Firstly, one can remark the existence of several types of reservoir or cistern, which correspond more or less to the types already identified during our study.

Secondly, several networks of channels and cisterns are superimposed and interwoven, from the deepest at 12.4 metres beneath the pavement of Fouad Street (1927 level) to the shallowest at 4 metres deep. Between these two levels one can remark several other intermediary levels.

As at Pompeii, it can be seen that cisterns that were at one moment connected to the supply network could subsequently function independently after the destruction or blocking of the supply channels, and inversely they could be reconnected to newly created supply channels.

For the Graeco-Roman period, the hypothesis elaborated would suggest that there was a dynamic system bringing water from the canal, after it had been filtered, right into the town. This water circulated through the numerous underground channels and supplied large volume reservoirs, as well as, perhaps, smaller storage volumes, as could have been the case for certain cisterns designed for domestic use. As the ground level of the town rose, and even if the initial network still existed and functioned, at least in part, it was necessary to complement the existing system with a new network established at a level that was more easily accessible from the surface. Furthermore, at known points on the line of the lower network and where the upper system past directly above, lifting mechanisms, such as saqias, were installed, in order to raise the water from the lower to higher level.
For what reasons were the Arabs obliged to create a new supply and storage system? This is one of the points that we are presently trying to address through explorations on site.