One has already been mentioned in connection with the possible ruins of a temple to Magna Mater, which were found in 1749. Among the Roman remains was an inscription to Mercury that was later used as construction material and can be seen today on an outer wall at the Travessa do Almada, near the Magdalena’s church. It is badly preserved and the text is practically gone, worn out by the elements and human action; it also appears to be broken, with part of the right side missing (and portions of the inscription with it). As it is, it reads: MERCVR… CAESA… AVGVST… C. IVLIVS F. IV… PERMISS V. DEC… DEDIT. F…, which allows us to assume it was dedicated to Mercury (and Augustus?) by a Caius Julius.
A clearer inscription appears to have been found during the 18th century near the Sun Gate, the eastern doorway in the 10th century Moorish walls. It read: MERCVRIO AVG. SACRVM C. IVLIVS CIVLII III AVGVSTALIS D. D., which has been translated as “Sacred to Mercury Augustus. The augustal Julius Catulius (?) gave and dedicated” (Vieira da Silva 1944, 188; n. 78). It is not certain if this particular votive inscription was near its original location. There are no traces of a temple in the area, but the 10th century gate may been built on an older road, if not originally a Roman doorway itself. And this can be deduced not by archaeological traces, but by the terrain itself, which today is still pretty steep in the eastern part of the old city. Back in the 12th century, during the siege of 1147, Flemish and German crusaders managed to bring down part of the wall near the Sun Gate, but the terrain proved too hard and they failed to enter the city (De Expugnatione Lyxbonensi 17). So that particular slope of the castle hill limits the possible gates and defensive lines, which in turn makes it possible – even if not certain – that the second votive inscription to Mercury may have been close to a road or pathway.
The third one may also have been found close to its original location. In 1940, during the castle’s (re)construction, a slightly damaged inscription was found, with the text MERCVRIO …HORIALI SACR. MLVIII.TVLA. Vieira da Silva assumed the missing letter in the second word to be a T, hence Thoriali, and quickly made a link with Thor (1944, 95). More realistically, José d’ Encarnação suggested a C, so Choriali from Latin cohors (1975, 234). If it’s a reference to the military cohorts, then it’s a votive inscription to a Mercury protector of the troops and could naturally be found on the fortified hill top. If, however, it means a courtyard or a fenced area for cattle, the inscription may have come from elsewhere to be used as construction material for the castle. It’s currently on display at the city museum.