Theodosius, on whom devolved the task of reorganizing the troops of the Eastern empire, appears to have appreciated to its fullest extent the military meaning of the fight of Adrianople. Abandoning the old Roman theory of war, he decided that the cavalry must in future compose the most important part of the imperial army. To provide himself with a sufficient force of horsemen, he was driven to a measure destined to sever all continuity between the military organization of the fourth and that of the fifth century. He did not, like Constantine, raise new corps, but began to enlist wholesale every Teutonic chief whom he could bribe to enter his service. The war-bands which followed these princes were not incorporated with the national troops; they obeyed their immediate commanders alone, and were strangers to the discipline of the Roman army. Yet to them was practically entrusted the fate of the empire; since they formed the most efficient division of the imperial forces. From the time of Theodosius the prince had to rely for the maintenance of order in the Roman world merely on the amount of loyalty which a constant stream of titles and honours could win from the commanders of the ‘Foederati.’ Only six years after Adrianople there were already 40,000 Gothic and other German horsemen serving under their own chiefs in the army of the East. The native troops sunk at once to an inferior position in the eyes of Roman generals, and the justice of their decision was verified a few years later when Theodosius’ German mercenaries won for him the two well-contested battles which crushed the usurper Magnus Maximus and his son Victor. On both those occasions, the Roman infantry of the West, those Gallic legions who had always been considered the best footmen in the world, were finally ridden down by the Teutonic cavalry who followed the standard of the legitimate emperor[13]. A picture of the state of the imperial army in the Western provinces, drawn precisely at this period, has been preserved for us in the work of Vegetius, a writer whose treatise would be of far greater value had he refrained from the attempt to identify the organization of his own day with that of the first century, by the use of the same words for entirely different things. In drawing inferences from his statements, it has also to be remembered that he frequently gives the ideal military forms of his imagination, instead of those which really existed in his day. For example, his legion is made to consist of 6000 men, while we know that in the end of the fourth century its establishment did not exceed 1500. His work is dedicated to one of the emperors who bore the name of Valentinian, probably to the second, as (in spite of Gibbon’s arguments in favour of Valentinian III) the relations of the various arms to each other and the character of their organization point to a date prior to the commencement of the fifth century.