Churches like those described in the last chapter, we have every reason to believe, were constructed before the Anglo-Norman invasion of this kingdom. How long the style continued is a matter of uncertainty. We find in the process of architectural development that the horizontal lintel appears gradually to have given place to the semicircular arch-head. The high-pitched roof becomes flattened, the walls lose much of their massive stone-work character and are generally higher, and cement appears to have been used. The windows exhibit a slight recess, or a chamfer, upon the exterior, and are of greater size; a small bead-moulding is occasionally found extending round an arch upon the interior. As the style advances, the sides of the doorways, though still inclined, become cut into a series of recesses, the angles of which are slightly rounded off. The addition of a slight moulding--at first a mere incision--upon the piers, would seem to have suggested pillars. Chevron and other decorations, which in England are considered to indicate the Norman period, are commonly found; but they are generally simple lines cut upon the face and soffit of the arch. Pediments appear; and the various mouldings and other details of doorways and windows become rich and striking, and, in some respects, bear considerable analogy to true Norman work. The capitals frequently represent human heads, the hair of which is interlaced with snake-like animals. To this style of architecture, which has its own distinctive characteristics, the term ‘Hiberno- or Irish-Romanesque’ has been applied. The churches, as a rule, are small in size and simple in plan, after the manner of the earlier buildings. Horizontal forms, so strong a feature in the old, are combined with the rounded forms in the new. The decoration contains elements distinctly belonging to that ornamentation which is so striking a feature of the metal and manuscript work of the period. This transition can be traced to the beginning of the eleventh century, but was not fully developed for a century later, and would lead, as Petrie says, ‘with every appearance of probability, to the conclusion that such architecture existed here previously even to the Norman Conquest of England.’[124] * * * * *