[121: Tacitus, _Ann._ ii. 64, says this expressly. Of free Thracians, viewed from the Roman stand-point, there were at that time none; but the Thracian mountains, and especially the Rhodope of the Bessi, maintained even in the state of peace an attitude as regards the princes installed by Rome, that could hardly be designated as subjection; they acknowledged the king doubtless, but obeyed him, as Tacitus says (_l.c._ and iv. 46, 51), only when it suited them.] [122: We have still a Greek epigram, dedicated to Cotys by Antipater of Thessalonica (_Anthol. Planud._ iv. 75), the same poet who celebrated also the conqueror of the Thracians, Piso (p. 24), and a Latin epistle in verse addressed to Cotys by Ovid (_ex Ponto_, ii. 9).] [123: It is one of the most seriously felt blanks of the Roman imperial history that the standing quarters of the two legions, which formed under the Julio-Claudian emperors the garrison of Moesia, the 4th Scythica and the 5th Macedonica (at least these were stationed there in the year 33; _C. I. L._ iii. 1698) cannot hitherto be pointed out with certainty. Probably they were Viminacium and Singidunum in what was afterwards upper Moesia. Among the legion-camps of lower Moesia, of which that of Troesmis in particular has numerous monuments to show, none appear to be older than Hadrian's time; the remains of the upper-Moesian are hitherto so scanty that they at least do not hinder our carrying back their origin a century further. When the king of Thrace in the year 18 takes arms against the Bastarnae and Scythians (Tacitus, _Ann._ ii. 65), this could not have been put forward even as a pretext, had lower-Moesian legionary camps been already at that time in existence. This very narrative shows that the warlike power of this vassal-prince was not inconsiderable, and that the setting aside of an uncompliant king of Thrace demanded caution.]