F. 5.--This story, as told by Constantius, and confirmed by both Nennius and Bede, incidentally furnishes us with something of a key to the main difficulty in accepting the widely-spread Romano-British Christianity to which the foregoing citations testify. What, it is asked, has become of all the Romano-British churches? Why are no traces of them found amongst the abundant Roman remains all over the land? That they were the special objects of destruction at the Saxon invasion we learn from Gildas. But this does not account for their very foundations having disappeared; yet at Silchester[432] alone have modern excavations unearthed any even approximately certain example of them. Where are all the rest? F. 6.--The question is partly answered when we read that the soldiers of Germanus had erected in their camp a church of wattle, and that such was the usual material of which, even as late as 446, British churches were built (as at Glastonbury). Seldom indeed would such leave any trace behind them; and thus the country churches of Roman Britain would be sought in vain by excavators. In the towns, however, stone or brick would assuredly be used, and to account for the paucity of ecclesiastical ruins three answers may be suggested. F. 7.--First, the number of continuously unoccupied Romano-British cities is very small indeed. Except at Silchester, Anderida, and Uriconium, almost every one has become an English town. But when this took place early in the English settlement of the land, the ruins of the Romano-British churches would still be clearly traceable at the conversion of the English, and would be rebuilt (as St. Martin's at Canterbury was in all probability rebuilt)[433] for the use of English Christianity, the old material[434] being worked up into the new edifices. It is probable that many of our churches thus stand on the very spot where the Romano-British churches stood of old. But this very fact would obliterate the remains of these churches. F. 8.--Secondly, it is very possible that many of the heathen temples may, after the edict of Theodosius (A.D. 392), have been turned into churches (like the Pantheon at Rome), so that _their_ remains may mark ecclesiastical sites. There are reasons for believing that in various places, such as St. Paul's, London, St. Peter's, Cambridge, and St. Mary's, Ribchester, Christian worship did actually thus succeed Pagan on the same site. F. 9.--Thirdly, as Lanciani points out, the earliest Christian churches were simply the ordinary dwelling-houses of such wealthier converts as were willing to permit meetings for worship beneath their roof, which in time became formally consecrated to that purpose. Such a dwelling-house usually consisted of an oblong central hall, with a pillared colonnade, opening into a roofed cloister or peristyle on either side, at one end into a smaller guest-room [_tablinum_], at the other into the porch of entry. The whole was arranged thus: