Dörpfeld has a theory about the origin of the Roman stage, which he brings forward as a strong argument in favour of his other views. According to Vitruvius the Roman stage was developed out of the Greek. The difference in size was due to the following reasons. The Romans preferred to give up the orchestra to the spectators and to transfer all performances to the stage. It was necessary, therefore, to deepen the stage, in order to find room for the additional performers. It was also necessary to lower it, in order to allow the spectators in the orchestra to have a clear view.[505] Dörpfeld says that this account of the matter is erroneous. According to his theory the Roman stage was discovered by accident rather than by design. The Romans, when they first began to adapt the Greek theatre to their own purposes, found the orchestra too large, and consequently divided it in two. The half nearest the auditorium they dug out to a depth of five feet, and placed spectators there. The other half they used for theatrical and other performances, just as it had been used by the Greeks. In this way they found that they had got what was practically a stage five feet high; and for the future, instead of digging out the nearer half of the orchestra, they started on the level, and built a raised stage. The Roman stage therefore represents, not the Greek proscenium, but the further half of the Greek orchestra; and this fact proves that it was in the orchestra that the Greek actors performed.[506] This theory is no doubt extremely ingenious. But unfortunately it appears to be inconsistent with the facts of the case. If it was true, we should expect to find the stage in all Roman theatres occupying the site of one half of the Greek orchestra, and the back of the Roman stage corresponding to the front of the Greek proscenium. Now in the normal Roman theatre this is more or less the case. The Romans eventually reduced their orchestra to a semicircle, and brought their stage forward to the position described by Dörpfeld. But the Graeco-Roman theatres of Asia Minor, to which we have already referred, fail entirely to correspond to his hypothesis. These theatres were among the earliest to be built in the Roman fashion, and might therefore be expected, more than any others, to exemplify the process of transition which he describes. But what do we find? We find that the stage, so far from occupying one half of the orchestra, stands in exactly the same position as the old Greek proscenium. The orchestra in these theatres still forms nearly a complete circle. The stage is deepened by pushing the back-scene more into the rear. Further than this, the height of the stage is not five feet, as it ought to be, but from eight to nine feet.[507] These examples seem to prove that Vitruvius is more correct than Dörpfeld in his view of the matter; and that the Roman stage was really a modification of the Greek. When we find in these Asia Minor