At Troesmis we conducted pioneering excavations in order to open up a new archaeological research horizon in the area that bore an exceptional significance during the Roman period and later, as resulted from the data on the few archaeological campaigns and from the ancient sources. The measuring profiles were oriented N 230 W so that they should not be parallel to any of the N – S, E – W, NW – SE or NE – SW directions, that characterize the ancient constructions. The general magnetometric map shows a very busy geomagnetic field, with unusually high values in comparison with other archaeological sites excavated by us in Dobrudja. The explanation must be that at Troesmis most constructions are of brick or granite, both with relatively high magnetic sensitivity, while at Histria, for instance, the ancient constructions are made of green schist, with a lower magnetic sensitivity than that of granite or fired clay, and at Adamclisi they are of limestone. The magnetometric map filtered by a high-pass filter sheds light on this issue, because it indicates the tracks of some linear-shaped constructions, situated under the level of the current ground, as the position of these constructions above ground is very close to that of the anomalies in question. Here we do not single out these constructions. They clearly appear on almost the entire surface of the map, under the form of some linear tendencies, oriented especially to the NW – SE and NE – SW. The excavations of the current campaign focused on the west area of the thermae, situated in praetentura dextra. We carried out nine cassettes, thus clearing a 250 m2 area. The excavation into room L was finished, the largest part of rooms M and N and partially the frigidarium I were excavated. Two praefurnia situated to the west of rooms L and M were excavated.