^{page 7 note b} Leland's Itin. edit. 1767, vol. vi. p. 139.

^{page 8 note e} Compare, with regard to this sea-marsh, William of Worcester, p. 147, (who states the width of the present outlet of the Severn between Aust Cliff and Chapel Hock, now a mile, to have been only a sling's throw at the time of his visit,) with the Map of Coal Districts, Geological Transactions, i. N. S. pl. 38. See also an account of this level in Seyer's Bristol, i. 138.

^{page 8 note d} In waiving this subject, it may, nevertheless, be desirable to notice some indications of Roman settlement at Stroat in Tidenham, and of communication between Stroat and Oldbury on Severn, which have escaped the numerous commentators on the 14th Iter of Antonine, though certainly reconcileable with it, (as transposed by Gale,) with one variation only, namely, the substitution of XI for VI in the distance between Abone, so transposed, and Aquae Sulis.

Under this arrangement, the distances (if measured direct, which may approach to balancing the differences between English and Roman miles) would be found to place “Trajectus” and “Abona” precisely at the positions of Stroat in Tidenham, and Knowle Camp in Almondsbury; and the intermediate “Sabrina,” (resting on the more apocryphal authority of Bertram's Richard of Cirencester,) would precisely coincide with Oldbury on Severn. All this, nevertheless, may be accidental coincidence.

Ancient roads (Roman or otherwise) are said to connect these points; but the present observations relate only to a possible Roman position, or adopted position, at Stroat. This village lies on that vicinal road from Venta to Glevum which has been noticed above, and which Leman, in his Commentary on Roman Roads (Hatcher's Richard of Cirencester, p. 114), has included in his Ryknield Street. Its distances are, between eight and nine miles from Caerwent, and four from the Roman camp and temple at Lidney.

At this point of Stroat, on a rising ground to the south-west of the village, are vestiges of excavations and earthworks, much reduced and altered by agricultural operations, and of unknown origin, which, combined with the former marshy banks of the Severn, would inclose a space of about ten acres, nearly oblong, but somewhat rounded on the south-west side, and which it is difficult to refer to any thing but military purposes. The Gaerston Hill farm, adjoining, evidently derives its name from them (as ancient remains of unknown origin), but the appellations of “Street” preserved in the name of the village, and of the “Oldbury Field” (situated within the traces of earthworks), point, as far as they go, to Roman origin. This name of “the Oldbury Field,” “in Strote,” is authenticated by the original Court Rolls of the manor of Waldings in 1614, and relates to a slight elevation, which the former earthworks, if continued, would inclose, and which lies between two brooks which fall, at a short distance, into the Severn at Horspill and Waldingspill adjacent, near a “Hoar Stone” of unusually large dimensions. A military station in this position would defend the narrow pass between the Chase Hills and the Severn, and also the landing-place of an irregular ferry, still used for occasional communication between Stroat and the opposite village of Oldbury on Severn three miles distant. A corresponding ferry of the Wye, at Tintern, is guarded by the Roman Camp of Madgetts, and there are intermediate earthworks (“the Bulwarks”), which seem to have been raised to command a later road, (passing between Bowlash Hill and Ashwell Grange,) but, at the early period under consideration, these would be useless, and natural woods and escarpments, with a military position on each river, would suffice to check all assailants advancing from the north towards the passages, or Caerwent.

But there are other lines defending the area thus inclosed, namely, the lines of Offa's Dyke (afterwards discussed), which skirt the Wye in Tidenham; and which lines Mr. Fosbroke (varying from received opinion) has considered “as communications between the Roman Camps thrown out to check the Silures,” which must have “guarded the Trajectus in a most powerful force, and almost invincibly have protected the passage of the Severn, through the numerous garrisons which could have been collected against an enemy within a very few hours.” (Gentleman's Magazine, vol. iii. New Series, 491.) A swbstructure by an earlier prince or nation is perfectly reconcileable with Offa's subsequent adaptation of these lines to his purposes; and the preceding observations tend, in some points, rather to confirm, than otherwise, this theory of a vast advanced position between the rivers. But it must be remembered that no Roman remains have been discovered here (except the uninscribed altar afterwards mentioned, from Madgetts), and the difficulties arising from such want of authentication increase in proportion to the scale of the settlement or station which requires it. A Roman position defended on the sides towards the Wye by the lines attributed traditionally to Offa, and on the northern side by the earthworks at Madgetts and the Chase Hills, would be nearly co-extensive with the entire parish of Tidenham, which contains about six thousand acres,

^{page 10 note e} Wood Chester, p, 18, and map, ibidem.

^{page 10 note f} Monmouthsh. vol. i. p. 2* and p. 21*.

^{page 10 note g} Memoirs of Bristol, vol. i. p. 78.

^{page 11 note h} It would, of course, pass easily also directly across the Wye to the present site of Chepstow, but the space within the walls appears to have been avoided by both the trackways. It must also be observed that there would be no accessible place of debarkation directly opposite to “the Bulwarks,” or to the south of this Pill, between it and Ewen's Rock, (which is about a mile lower down,) on account of the intervening alluvial sea-marsh. As late as 1651, an inquisition was taken in Cromwell's Court of Survey for his manor of “Tydnam,” to define the road to this Ewen's Rock, which was probably at that time a general landing place within the Wye, and at the present day the marsh would be impassable without continual drainage.

^{page 11 note i} The fosse of this camp (which is in itself a most picturesque scene, from its connexion with the Cliffs and the Wye, and commands the varied prospects at the confluence of the two rivers and the Bristol Channel) has lately been cleared of brushwood and other obstructions under the direction of the Bishop of Llandaff, and forms a most interesting addition to his grounds at Hardwick.

^{page 12 note k} I apprehend that it diverged from the present line of turnpike road, or its neighbourhood, on the southern side of the brook at Pwll-Meyric, and passed up the deep valley between Crick and Runston. Beyond Crick its line has long been noticed by Coxe and others.

^{page 12 note l} Coxe, who was unable to find them (Monmouthshire, ii. 364), must have sought them when covered by the tide, or at a wrong point. In the notes on the “Cygnea Cantio,” Leland cites the note of a nameless commentator on some verses of Alexander Necham (Abbot of Cirencester), which mention the bridge on the Strata Julia, “quod vulgo Strigolium dicitur,” (Itin. ix. 101), but the passage is unworthy of any particular consideration.

^{page 13 note m} Gale, XV. Scriptores, i. 194.

^{page 13 note n} Britannia, edit. 1590, p. 499.

^{page 13 note o} I believe that the Cambrian authorities agree with the English Chronicles in thus referring both lines to Offa,; but it is not intended to enter on their possibly consecutive construction, the alleged breaking down of a portion of the line, the Mercian colonization, or any thing beyond the statements requisite to lead the traced line of demarcation southwards to the point where tradition connects it with personally investigated remains. The reader may find the general statements in Caradoc's Welsh Chronicle, edit. 1584, p. 19, and a story which may refer to some portion of the outer line, in the additions to Matt. Paris. (Vita Offæ Secundi.) Edit. Wats, p. 47.

^{page 13 note p} Pennant's Wales, 4to edit. ii. 273.

^{page 14 note q} It is uncertain whether the Wye itself was the boundary through the southern part of Herefordshire, or whether a continuation of the dyke once existed here, in a course which Sir S. R. Meyrick has attempted to trace by indications of local names, in a memoir preserved in the Cambrian Magazine, and also in the Gentleman's Magazine, vol. ciii. part i. 504. See also Sir R. C. Hoare's remarks, Girald. Camb. vol. ii. p. 394.

^{page 14 note r} Mr. Fosbroke's personal examination of the dyke through these coppices is given in Gentleman's Magazine, vol. cii. part ii. p. 501, and he has other communications in vol. ci. part ii. and in vol. iii. new series, in the latter of which he appears to doubt the fact of the dyke, which he had traced along the Wye between Monmouth and Tidenham, being the genuine Offa's Dyke, chiefly from its not pursuing a direction conformable with precise points of the compass. Sir Samuel Meyrick's observation on this objection is, that Roman roads through conquered countries varied little from precise directions, but that Offa, marking “the boundary of his kingdom, extending much further west in some places than in others, could not avoid giving to his work an irregular appearance.” A further conjecture of Mr. Fosbroke's, as to these lines being Roman, is noticed in page 9. No objection is offered to a theory of earlier works as a possible substructure, and to an adaptation of it, by the Mercian King, to his later purposes. The Dyke would not be less entitled to be considered Offa's Dyke from being, in some places, re-constructed on earlier foundations.

^{page 16 note s} Worthines of Wales, edit. 1776, p. 104.

^{page 16 note t} Corbet, p. 136, in Bibl. Gloucestrensis.

^{page 17 note v} In his Introduction prefixed to the Gloucestershire Civil War Tracts, p. xcvii.

^{page 17 note w} Mr. Jennings, Custom House Officer, Chepstow. See Tour in Monmonthshire, p. 366.

^{page 17 note x} Giraldus Cambrensis, ii. 395.

^{page 17 note y} The precise point of Massey's attack was at Buttinton, where a part of the Dyke is sloped on the Tidenham side into an irregular glacis. Cannon shot are occasionally dug up in a line from this point down to the moated site of Badamscourt, which was probably burnt by Wintour before those battles, as the remains of the walls shew marks of fire, and other portions of them were found lying within the moat (when lately cleared) in entire masses, as having fallen in the confusion of such destruction. No trace of the more modern works added by Rupert and Wintour remains. A large portion consisted of palisades.

^{page 18 note z} See Ingram's Saxon Chronicle, p. 117, and Matthew of Westminster, p. 179, edit. 1601. The description of the latter is as follows:

895. “Eodem tempore Hasteinus iniquus et ceteri paganiquos exercitus Regis AElfredi de Beamfleote fugaverat, disposuerunt transire ad concives suos, qui in occiduis partibus Angliæ habitabant. Arrepto igitur clam itinere per provinciam Merciorum, ad villam quandam super flumen Sabrinæ sitam, Buttingdune appellatum pervenerunt, ubi a confratribus suis reverenter sunt recepti, atq; in oppidum quod ibidem construxerant reverenter sunt recepti. Cumque hoc Regi Ælfrido nuntiatum fuisset, congregavit exercitum quasi invincibilem, et veniens ad oppidum prefatum, quod Sabrinœ fluctibus erat undique perfusum, paganos tam navali quam terrestri exercitu circumcinxit. At hostes post diutinam obsidionem, victu sibi deficiente, equos suos novissime devorabunt. Tandem, omnibus consumptis, ab oppido necessitate compulsi exierunt, contra exercitum qui erat en orientali parte fluminis pugnaturi, ubi ex parte Regis in prima congressime cecedit Ordeinus minister ejusdem regis et multi alii cum eo, sed Christiani demum prevalentes in fugam adversarios compulerunt. Quos fideles sine pietate insequentes, multos in undis submergebant, et nonnullos gladiis detruncabant.”

^{page 18 note a} The width of the Severn here must have been very inconsiderable at this period, as William of Worcester states, in 1453, that Aust Cliff was within a stone's throw (of course from a sling) from that Rock which is still joined to Beachley on the opposite side at low water—“Charstone Rok est ita magna rupis sicut rupis Sancti Tiriaci, distans a firma terrâ de Austclyff per jactum lapidis.” For this he quotes “Informacio Will'mi Powell de Tynterne;” but this ancient preserver of measurements might have quoted from personal knowledge, for he had just visited Tintern Abbey, “de Austclyff per aquam usque Chepstow navigando.” p. 144.

^{page 19 note b} Mr. Harris, looking for the Roman Trajectus at this point in 1763, says, “by all the inquiry I could make, there do not appear any visible traces of a work of this kind at Tidenham or Beachley.”—Archaeologia, vol. II. p. 2.

^{page 19 note c} See Mr. Dawson Turner's remarks on the leaden font of Bourg Achard, in Normandy, ii. 97; and Mr. Gough's observations on ancient fonts, Archaeologia, vol. X. 187; and also Canon 81, directing “a font of stone” in every church (according to a former Constitution), “in which only font the minister shall baptize publicly.”

^{page 20 note d} Amongst much mutilated glass in the south windows remain the arms and monogram of Sir John ah Adam, Baron of Beverstone, heir male of Herbert of Llanlowell, a warrior in the Scotch expeditions of Edward I., and husband of the rich heiress of the houses of de Gournay, Gant, and Berkeley of Were. The glass is of interest, as the seal of this Baron is wanting among those attached to the Letter of the Barons to the Pope engraved in the Vetusta Monumenta. Its position at Tidenham is accounted for by the Baron's hereditary possession of the Herbert manors of Bettesley and Gorse, with the mansion of Badamscourt in this parish, which (as well as Llanlowell) were enjoyed by his descendants, long after the Gournay estates had been dissipated by his son. The arms are, Argent, on a cross gules five mullets or.

^{page 20 note e} As the designation of the contiguous woods of Banagor may refer to the British name for places of education and religion, or to Bann-y-gawr, collis gigantis, I leave the inference to Cambrian Antiquaries.

^{page 21 note e} Hermes Britannicus, London, 1828, p. 79 and 105.

^{page 21 note e} Hermes Britannicus, London, 1828, p. 79 and 105.

^{page 21 note f} It must be distinguished from the “Lancoit” of another early document, “the Book of Llandaff,” (Mon. Ang. vi. 1225,) which follows the lands granted by British Princes to that See, extending from Matherne to Portcasseg immediately opposite to Llancaut. The site of this other place of kindred name was in Llyswen parish, Brecknockshire.

^{page 21 note g} See the late edition of the Monasticon, vol. ii. 256, 264: and Kemble's Codex Diplomaticus Ævi Saxonici, ii p. 327.

^{page 22 note h} Ord. Vit. p. 704. Simeon Dunelm. p. 222.

^{page 22 note i} Some further particulars respecting it may however be requisite, with respect to what is afterwards stated of the Counts of Eu, the hundred of Twiferde, which identifies the locality of this part of Tidenham, and the connexion of the whole subject with the boundaries cited afterwards to prove the position of “Strigul.” The forfeiture of the English lordships of William Count of Eu is mentioned in the text, in which Odelaveston or Woolastone accompanied Tidenham. In 1073, after the defeat of Roger Earl of Hereford at Fakedun (Fakenham?) by William Earl Warren and Richard Fitz Gilbert, the Earl of Hereford's portion of Tidenham was also forfeited. In addition to Strigul Marchership, these two forfeited portions of Stigand's lordship of Tidenham, with Woolaston, were afterwards vested in the Clares Earls of Pembroke and Marshalls of England, (possibly, and in part by a remunerating grant from the Crown to Richard Fitz Gilbert himself,) and constituted “the Earl Marshall's Liberty,” extending from Strigul Bridge to Cone in Alvington. This, by the 27 Hen. VIII. is considered part of the Marches and annexed to Gloucestershire, but in the Hundred Rolls (i. 181), it is distinctly and truly stated to have been formerly part of the county (though then withdrawn from it by some means unknown), and to have been included in Twiferde hundred to the time of King John. This mode of mixing up Saxon hundreds with lordships marcher, and with what was obtained (to use the words of the very inaccurate Tintern Chronicler) “licentiâ conquerendi super Wallenses” (Mon. Ang. v. 270), appears to be an anomaly.

^{page 23 note i} Mon. Ang. ii. 256.

^{page 23 note k} Vincent's Discoverie of Errors, p. 234.

^{page 24 note l} Archaeologia, vol. XXIV. p. 87.

^{page 24 note m} Ibid, pp, 32, 33.

^{page 25 note n} The confusion of opinions on this subject has been great. Sir R. Atkins, misled by the juxta-position of Gloucester and Estrighoiel in Domesday (Hist, of Glost. p. 45), unaccountably translates “Castellum Estrighoiel,” as “The Castle of Eastbridge Hotel in Gloucester.” Gale fancies that he sees “Strata Julia” in it, and Humphrey Lhwyd considers it “somewhat neare to Siluria.” (Twyne's Translation, fo. 81 b.) Williams, in his Monmouthshire (p. 140), not being aware of the error of the Tintern Chronicler (Mon. Ang. v. 269, no. iv.) who confounds the two principal Norman grantees in Tidenham (William Fitz-Osberneof Bretteville and Hereford, and William Fitz-Robert, of Eu), and the two successive lines of the Norman Counts of Eu, thinks that the names of both castles are derived from “Castrum Ogie,” as having both been held by “Clare, Lord of Ogie.” It is true that the Clares were male descendants of the eldest line of the Counts of Eu, and that after the two forfeitures already mentioned (as explained in a previous note) they obtained both the portion of Tidenham which had belonged to their kinsman William Count of Eu (representative of the second line), and that of Roger Earl of Hereford, whose rebellion was put down by Richard Fitz-Gilbert, surnamed de Tonebrugge, and the great ancestor of this Norman house; but Clare of Strigul and Pembroke never possessed the title of Count of “Ogie,” or Eu, nor were the two Monmouthshire castles ever possessed by any one holding such title. This subject may be traced in Duchesne's Scriptores Hist. Normann., to which the appended pedigrees give easy reference.

Apparent whimsical corruptions of this British name, “Traigyl,” or “Treigl,” may he found in the Wye nearer its confluence with the Severn, in the field at that point, and in the name of the rocky island and anchorite's chapel at the junction of the rivers. William of Worcester sometimes terms this, the “Rok Seynt Tryacle,” and sometimes Latinizes it into the “Capella Sancti Teriaci Anachoritæ.” In Saxton's map it appears as “St. Treacle Chapel.” The Valor of 26 Hen. VIII. has “Capella Sti Triaci,” and the ruins are often called St. Tecla's (euphoniæ gratiâ), but the natives and mariners uniformly use the absurd corruption given by Saxton. In this derivation, as in the other, I rest on the authority of a native antiquary. Another prefers Trigle, habitationis locus, but the Troggy is a stream of the Forest, which takes the name of Nedern, when it approaches the habitations of man, and the Station of Venta.

^{page 26 note o} If by this expression Camden intended to affirm the destruction of the still existing walls, he must obviously have written on this point without local communication.

^{page 27 note q} Itinerary, vol. vi. fo. 23, and ix. 101.

^{page 27 note r} As a proof of the names being occasionally used indifferently, may be cited, on one side, from Rot. Parl. vi. p. 207, 22 Edw. IV. the saving to William Erie of Huntyndon, of the “Castelles, Lordships and Manors of Chepstow and Gower;” and ibid. pp. 292–3, the petition of the coheirs of Charles respecting proceedings connected with the transfer from the Duke of Norfolk to William Lord Herbert of the Lordships, Manors, &c. called Gower Lands and Shepstow; and, on the other hand, in Twyne's Breviary 1573 (translated from Humphrey Lhwyd's Commentarioli Descriptionis Britannicæ), “Chepstow, a fine market town in Wenta, before a few years passed, was called Strigulia.”

^{page 27 note s} Rot. Pat. 42 b.

^{page 27 note t} Rot. Orig. i. 154.

^{page 27 note u} XV. Script, p. 194.

^{page 27 note x} Domesday, i, 162.

^{page 27 note y} It may be excusable to diverge from the immediate object under discussion, to observe that there is nevertheless much relating to ancient castles (though not so named) within the Domesday Survey of the district apparently attached to Estrighoiel, which stretched to the Usk, and constituted the germ of the future marchership. Among these will be found “Caroen,” (the ruins of the Roman Venta Silurum,) and three “Hardvices,” or (fortified?) granges within the part of such district distinctively called “Wales.” These are, Lamecare, Poteschevit, and Dinan, viz. Llanvair, Portskewit, and Dinham. At the first and last of these ruins of Norman castles still exist, and at Portskewit there are earthworks near the church, and also very strong ones at Sudbrook (which has been considered a Roman Station), but it is impossible to connect either of these, precisely, with the “Hardwick” of Domesday, or the earlier palace of Harold, burnt in 1065 by the Gwentian Prince, Caradoc ap Gruffyth, mentioned as “Rex Caraduech” in Domesday. Caldecot is also named, but the British Episcopal Seat of Matherne, and almost total ruin of the Bishopric of LlandafF, are unnoticed.

^{page 28 note y} Mon. Ang. vol. v. p. 268.

^{page 28 note z} MS. “Treatise of Lordships Marchers,” p. 138, abstracted in Appendix to Pennant's Wales, vol. ii. 4to edit. p. 443.

^{page 28 note a} An ancient translation of this charter is preserved in the Duke of Beaufort's office at Chepstow; the original is supposed to be lost.

^{page 29 note b} Neustria Pia, p. 600.

^{page 29 note c} Mon. Ang. iv. 597,

^{page 30 note c} The windows are of various ages, but none of the regular Norman form represented in some late engravings. In the writ to John de Monemue to take possession of the castles of Gilbert Marescall', late Earl of Pembroke (Pat. Fin. 25 Hen. III.) Strigoyll, Usk, and Karelioun are only noticed, and there is no existing part of Troggy that appears older than this period. The partition of these Lordships is supposed to have followed the death of Earl Anselm in 29 Hen. III. and the erection of a Castellet at this point would then become desirable to protect the new frontier, which soon afterwards was the scene of feudal warfare between the Earls Marshall, Lords of Strigul, and the Clares, Earls of Gloucester, as Lords of Usk.