In 1858--the year in which Archdeacon Trollope published the results of his researches--Capt. W. H. Mounsey drew attention to the description in a Welsh history book ("Drych y Prif Oesoedd," published in 1740) of a curious custom formerly prevalent among the Welsh shepherds. This custom consisted of cutting in the turf a figure in the form of a labyrinth, which they called Caerdroia, _i.e._ the walls, or citadel, of Troy. He also remarked that the herdsmen of Burgh and Rockcliffe "at the present day are in the habit of cutting this labyrinthine figure, which they also call 'the Walls of Troy.'" He drew the tentative conclusion that this name "would seem to be an after-thought of pure Cymric origin, suggested by the similarity between _Caerdroia_, the City of Troy, and _Caer y troiau_, the city of windings or turnings." A similar suggestion had already been made in the _Transactions of the Cymmrodorion Society_ in 1822, the writer ("Idrison") holding that the turf figures, and also those on the Knossian coins, had reference to the courses of the sun as conceived by ancient worshippers of that orb. Captain Mounsey was promptly answered by Dr. Trollope, who referred to the wide distribution of these devices throughout England and commented on their total absence from Brittany, where, if they were of ancient Cymric origin, one would have expected to find at least some trace of them. He also stated that they first received the name of "Troy-towns" in Tudor days, when "subtleties" of all kinds were in vogue, the term being used simply to indicate, by analogy with the Troy of legend, the difficulties to be overcome before the centre could be reached. Dr. Trollope gave it as his considered opinion that they were originally cut for penitential purposes by ecclesiastics, and this opinion has since his time remained practically unchallenged. In his memoir on the subject he reproduces a sketch showing the St. Anne's Hill maze with two gowned and kneeling persons in the act of performing a penitential circuit. Both the sketch and Dr. Trollope's conclusion are based on inference, however; there does not appear to be any direct evidence in the matter. The theory of an ecclesiastical origin of the turf mazes is chiefly supported by analogy with the continental church labyrinths which many of them so strongly resemble. Against the argument of their frequent proximity to an ecclesiastical site we may place that of their equally frequent proximity to known Roman remains and the fact that many of our old churches were founded on Roman sites.