Not connected with the subject of holy wells, but apparently forming isolated and local features peculiar to Wales, are the well-known oval or circular churchyards, enclosing churches which date from the Norman period. The churchyards are usually encompassed by a road, for which there is no obvious public requirement[261]. It has been conjectured that these roads represent ancient ramparts, which separate the churchyard from common ground, and this prosaic explanation may be the correct one. But one is obliged to notice another ray of light which comes from ancient custom. The Rev. E. Owen, who has described these churchyards, sees an analogy to the circle of stones in which religious ceremonies were performed by the Druids[262]--evidently he is referring to historic times. These circles, when prehistoric, are known to the archaeologist as “cromlechs”; the latter erections, from the fifth century onwards, were technically called “gorseddau” (_sing._ gorsedd). The Gorsedd consisted normally of a mound of earth and a circle of standing stones[263]. From denoting the place of assembly, and afterwards, “the Great Seat,” the word came to mean the “Assembly of Bards,” the chief member of which was throned on a “Chair,” or stone, which occupied the centre of the circle. So early as the ninth century, there was a separation of functions; hence we read of the gorsedd of the bards and the legislative gorsedd[264]. My friend, the Rev. J. W. Hayes, who has collected much curious lore respecting the gorseddau of later centuries, notices that, though the legislative gorsedd has now no political or judicial powers, but merely controls the bardic order, it has a successor, for all worthy aims, in the national Eisteddfod. The Eisteddfod has social and educational functions only, the Gorsedd, on the contrary, was an institution for the framing of laws. Even in the year 1910, however, the Eisteddfod was preceded each day by the Gorsedd proper. This slight description will enable us, in the next chapter, to approach closely to another side of our problem, but, for the present, it must be taken as illustrative of the supposition made by Mr Owen. From the fact that, at one village, Efenechtyd, Denbighshire, a part of the encircling road really occupies the ancient bed of a stream, Mr Owen has further considered that the “roads” were originally intended to be moats, and that they contained water. This seems to be mere speculation; a more plausible explanation--though, again, perhaps not the real one--is that the hollows formed portions of an old stockaded village. Or again, we may have here small ring earthworks belonging to the pre-Christian period, though not necessarily of a defensive character. One cannot avoid recalling Stonehenge and Stennis; the round churches of Northampton, Essex, Cambridge, and London; the round towers of many other churches; the favourite “broken ring” of Bronze Age barrows and Bronze Age ornaments; and the earthwork rings and circular