This is the third article in a line of preliminary reports on Croatian-French excavations in Salona, precisely in the Episcopal Centre - Oratorium A. The excavations of 2004 were not as long as the ones of the previous years. The work in the southern part of chamber I.2 continued. Excavations of 20020 layer composed of yellowish-brown soil with pottery fragments such as amphorae, tiles, roof tiles, fine dishes, animal bones, and small stones, were completed. Layer 20021 is identical regarding its composition except for lumps of mortar bearing stone prints - it is the matter of mortar that connected the wall. The described layers are obviously a result of disposal of construction waste, kitchen waste and discarded pottery fragments. During the excavations of 2002 and 2003 we established that the Oratorium building had been renovated in detail at least twice: small baths were constructed, and later on demolished. During the last year’s excavations in the chamber with a semicircular bench (I.1) in the trench under the pavement (trench A) we discovered late Roman coins, the most recent one among them minted between 375 and 392. We put a trench in the north-western corner of the chamber (trench B) with the aim to find out if the pavement in the western part of chamber I.1 had been made at the same time as in the eastern part. In the trench we found a wall in the north-south direction that preceded the construction of Oratorium A, and of the semicircular bench (I.134). In layers 10014 and 10016 fragments of late Roman pottery were excavated, among which a fragment of North African lamp and a fragment of the rim, neck and handles of amphora (Cat. Nos. 1 and 2) were selected. In an earlier research in and analysis of the technique and order of building the walls, it was established that the northern wall of the chamber with a semicircular bench (I.1) had been raised on the wall belonging to an earlier structure below the Oratorium A. The walls of the Oratorium A were not built exactly following the line of the former walls. The foundations of wall I.101 (the wall of the Oratorim A with butresses built against the northern side) are extremely weak, with small stones connected with little mortar. The discovery of the weak foundation structure of a part of wide wall with the butresses to support it should be linked with the purpose of the massive masoned structure (I. 901) which fills the space between the last two western butresses against the southern wall of the Oratorium A (I.103). This is a very firm structure, whose southern face was built of large stones, and it is completely filled with stones that are connected with lots of mortar. W. Gerber interpreted it as a prefurnium. This year we established that the south-eastern corner of the Oratorium A was built on the canal. Due to the poor foundation the southern wall cracked and a support was necessary to stabilize it on the southern side. The massive masoned structure was built to support the cracked wall. The builders of the structure with butresses seem to have raised a new object on the remains of the old one however, without thinking whether the old walls were still there or not, that is, paying no attention to the stability of the building. This might mean that the building of the Oratorium A was built “in a hurry”, possibly at a dangerous moment (it is situated close to the town walls and aquaeduct). This idea suggests that it could have been a “military” (defence? protection of aquaeduct?) object built on the former building in the corner of the town walls and near the aquaeduct. If this supposition is right, the original use of the building obviously changed with the time. The remains of the canal and walls, which must have belonged to the drainage system from the aquaeduct, were excavated in two trenches in the western yard (area I.9). The analysis of numerous (179 pieces of) mosaic fragments did not prove the original position of the mosaic. The article ends with a catalogue of the selection of ceramic finds from trench B (layers 10016 and 10014).