These are collected from a comparison of English examples, with the accounts of those still existing in Normandy. The nature of Norman fortresses is perhaps more definite than that of later erections, by reason of its greater simplicity; for in after times, military architecture was accommodated more to the conveniences of life, and a compromise was thus made as to the form and arrangement of castles. In various strong positions throughout this country we find, as in Normandy, many deeply entrenched earth-works, which, though attributed to the Danes, may in many instances be more correctly ascribed to the ninth or tenth century, or even a much later period. Instances among others occur at Builth, South Wales, where the two moats and the motte or keep-mound are in their original state, but with no remains of masonry; also at Longtown, at the foot of the Black Mountains, Herefordshire, where the motte and circular keep remain, and again at Kilpeck, in the same county, are extensive earth-works with the masonry destroyed. The origin of this castle is historically recorded, and the account given at length in Lewis's Kilpeck Church. I may mention a fourth, perhaps earlier, at Laughton-en-le-Morthen, near Doncaster, East Yorkshire, consisting of an outer ballium, and a lofty motte, but without any traces of masonry. Close adjacent is the church, exhibiting manifest traces of Saxon architecture. My note-books, if I had time to search them, would I doubt not present several other examples; the fortifications of Old Sarum are also an example. The classification of the different descriptions of earth-works, which are found in almost countless numbers throughout the country, has not yet been properly undertaken, although I think offering an interesting field for investigation, particularly if regarded in connection with similar remains in other countries. As regards our present subject—Norman fortresses—some assistance may be derived from a comparison with those of Normandy.
The earliest military works seem to have been derived from the forts erected by the Romans in their distant provinces. In the time of Justinian we read of castles strongly posted on commanding heights defended by precipices and deep ditches; with battlements, machicolations, portcullisses, and galleries in the thickness of the walls. The keeps of our Norman castles seem to be the probable representatives of the Roman prætoria. And here I may mention the singular occurrence
