APPENDIX.-No. I

COMMENTARY ON THE ITINERARY.

No people are so barbarous as to be totally destitute of the means of int nal communication; and in proportion as they become more civilized a have more intercourse with other nations, these means are augmented and facilitated. By the early accounts of the Britons it appears that they maintained a considerable foreign commerce, that they had formed towns or large communities, and used chariots for warlike, and undoubtedly for civil purposes. Hence it is evident that their internal communications must have been free and numerous. We need not therefore be surprised, if, after the lapse of so many centuries, marks of such British roads appear even at present to a careful observer, differing in many respects from the roads subsequently made by the Romans, and traversing the island in every direction.

These ancient ways may be distinguished from those made by the Romans by unequivocal marks.

1. They are not raised nor paved, nor always straight; but often wind along the tops or sides of the chains of hills which lie in their course.

II. They do not lead to Roman towns, or notice such towns, except when placed on the sites of British fortresses.

III. They are attended by tumuli like those of the Romans; but usually throw out branches, which, after running parallel for some miles, are reunited to the original stem.

When the Romans obtained a footing in this island, they directed all their operations, according to their practice, by military principles. They civilized indeed as they conquered, but conquest was their principal object. Hence, as each tribe was successively subdued, they fortified such primary posts as were best adapted to support their future operations, established secondary posts to secure their communications, and connected the whole by military ways. From local circumstances, and the principles of war, their primary posts were either at or near the sites of the British towns, or on the principal rivers. If therefore the British towns and trackways were suited to their purposes, they adopted them; if not, they constructed others. But both their towns and roads differed materially from those of the original inhabitants. The Romans in their towns or fortresses followed the system of their own castramentation, in like manner as in modern warfare the construction of permanent and temporary works is guided by the same general principles. These towns are of a regular figure, bounded by lines as straight as the shape of the ground will permit, generally square or oblong, and consisting commonly in a single wall and ditch, unless in positions peculiarly dangerous, or where local circumstances rendered additional defences necessary. On the contrary the British towns, which were occupied by the Romans, although irregularly shaped, still partake of their original figure.

Specimens of the first kind, or perfect Roman towns, may be seen in Colchester, Winchester, Caerleon, Caerwent, Castor near Norwich, and all

the military stations bordering on the wall of Severus. Of the latter, in Bath, Silchester, Kentchester, Canterbury, and other places.

Similar marks of difference between the original British trackways and the Roman roads appear in the Foss, and the Iknield Street ;-the latter, during the greater part of its course, keeping along the chain of hills which lay in its way, not leading decidedly to Roman towns, throwing out parallel branches, attended always with tumuli, still bearing its British name, and appearing from its direction to have been made for commercial purposes.

On the other hand the adopted roads, but more especially those made by the Romans themselves, are distinguished by peculiar marks. Posts or towns are placed on them at nearly regular distances, seldom exceeding twenty miles, the length of a single march, and also at the point where two roads intersect each other, or where several roads diverge. These roads are elevated with surprising labour to the height of ten feet, and sometimes even more, instances of which may be seen on the heath near Woodyates Inn in Dorsetshire, near Old Sarum on the side of Ford, in Chute Park, Wilts, between Ancaster and Lincoln, and still more remarkably on Bramham Moor, near Tadcaster in Yorkshire. They were formed of niaterials often brought from a considerable distance, such as chalk, pebbles, or gravel; and the most considerable are paved with stones, which are visible to this day. Tumuli also, which seem to have been the direction-posts of antiquity, attended their course, and occur in almost every instance where a road descends a hill, approaches a station, or throws off a branch. Another peculiarity of the Roman ways is their straight direction, from which they seldom deviate, except to avoid a rapid ascent or descent, to throw off another road, or to approach a station, which, from the circumstances before mentioned, had been fixed out of the general line. Of this there is a curious instance where the Foss, in approaching Cirencester from the north, meets the Akeman Street, bearing to the same point from the north-east, and evidently bends out of its course to join and enter the station with it.

Of many of the Roman roads, not only in England, but in the greater part of the Roman empire, an account has been preserved under the name of the Itinerary of Antoninus, which specifies the towns or stations on each road, and shows the distances between them. This record was long supposed to be a public directory or guide for the march of soldiers; but if this were the case, it is extremely confused and imperfect. It often omits in one Iter or journey towns which are directly in its course, and yet specifies them in another, as may be seen in the first, second, sixth, and eighth Iters. It traces the same road more than once, and passes unnoticed some of the most remarkable roads in the island, namely a great part of the Foss, and the whole of the Via Devana (a road from Colchester to Chester.) Hence this Itinerary has been more justly considered as the heads of a journal formed by some traveller or officer, who visited the different parts of the empire from business or duty; and, as Mr. Reynolds conjectures with great appearance of probability, in the suite of the emperor Adrian. In this light it may be considered as copious, and the advantages which it has afforded to the antiquary will be gratefully and universally acknowledged. Still, however, from the incoherence which appears in that part relating to our island, and from the mutilated copies which have been found, there is reason to imagine that the whole of this interesting record has not escaped the ravages of time.

Such an itinerary, but varying in many respects from that of Antonine, is one of the most important parts of the work now presented to the reader.

In fixing the sites of the towns specified in these Itineraries, our antiquaries have assumed the most unjustifiable latitude. The mere resemblance of a name was considered as a reason sufficient to outweigh all others; even the great Camden suffered himself to be misled by this resemblance, in fixing Ariconium at Kentchester, Camalodunum at Maldon, Bennavenna at Bensford, Pons Elii at Pont Eland, and Ad-Pontem at Paunton. The explanation of the names to suit the supposed situation has been another fruitful source of error; not only British and Latin, but Saxon, Greek, and even Hebrew, have been exhausted to discover significant appellations; and where one language was not sufficient, half a word has been borrowed from one language and half from another to support a favourite hypothesis.* The commentary now presented to the reader is founded on the following principles.

I. The vestiges of roads actually existing are taken as much as possible for guides; and the extremes or direction of each Iter, ascertained from two or more undoubted stations, or other unequivocal proofs.

II. In general, no place is regarded as the site of a Roman station, unless fixed Roman remains, such as buildings, baths, &c. are found at or near it; and unless it is situated on or near the line of a Roman road.

III. An exception has, however, been sometimes unavoidably made to this rule. After the Romans had established their power, and completed their system of internal communication, they undoubtedly lessened the number of their garrisons, to avoid either too great a division of their force, or to reduce that part of it which was necessarily stationary. Hence we have sometimes considered the direction of the road, and the general distance, as sufficient data for determining a station or stations, either when they were situated between two considerable fortified points, or when covered by others on every side; because it is probable such posts were merely temporary, and were dilapidated or demolished, even before the decline of the Roman power.

IV. In assigning a specific Roman name to a place, it has not been deemed sufficient that fixed antiquities or other equivalent evidence prove a town to have existed on the spot, unless the order of the names, and the distances marked in the Itinerary, justify the appellation.

V. Where the line of the Roman road is tolerably perfect, no station is sought far from it, except where the excess of the Itinerary over the real distance, or accurate measurement, affords sufficient authority for the deviation.

VI. The numbers which determine the distances being written in Roman numerals, which gave great latitude for errort and substitutions, recourse has been had to this rule.

On this subject it may not be improper to observe, that the name of Castor, Cester, or Chester, generally points out a Roman station; and Sarn, Street, Stane and Stone, (Strat, and Stan, when compounded) as generally show the course of a British or Roman way.

+For example these marks being the mutilated parts of numerals, might have been easily transformed by the copyist into IIIII. XIII. VIII.

Where the road still exists, the whole intermediate space between two stations already determined, has been examined to discover what places, from their relative distance, from their site, or the antiquities found in them, have the fairest claim to be considered as Roman posts; and to such places the names have been affixed according to the evidence afforded in the Itinerary.

After this development of the principles on which we have proceeded in our examination, it is necessary to add a few observations on the Roman mile, the standard of measurement used in compiling the Itineraries; because many difficulties in determining the stations arise from our uncertainty respecting its real length. It may indeed appear easy to ascertain this point, by a careful measurement of the space between two military columns, still existing on any known Roman road. But in Britain such an experiment has been hitherto impracticable; for the columns in our island have been so universally defaced or removed, that, far from two existing on the same road, only one has been found* whose original station is known with any degree of certainty. In France and Italy many of these columns still exist, and Danville has adduced three instances in Languedoc, in which the distances between them accurately measured amounted in one to 756, in another to 753, and in a third to 752 toises and two feet, The average 754 toises and two feet, seems to determine the length of the Roman mile with sufficient precision; and the result is confirmed by a comparison with the Roman foot, still preserved in the capitol; for the exact length of the miles between the military columns on the Appian way, in the neighbourhood of Rome, as measured by Bianchini, was 5010 of these Roman feet, which reduced to toises is 756 toises four feet and a half. From these results Danville estimates the Roman mile at 755 toises, or 1593 yards+ English measure.

Unfortunately this mensuration does not lessen the difficulties of the English antiquary; for the distance between any two of our known stations, if measured by this standard, disagrees in almost every instance with the numbers of the Itineraries. Different conjectures have been advanced to solve this difficulty. One, supported by the respectable authority of Horsley, is, that the Romans measured only the horizontal distance, without regarding the inequalities of the surface; or that the space between station and station was ascertained from maps accurately constructed. This idea receives some support from a fact acknowledged by every British antiquary, namely, that the Itinerary miles bear a regular proportion to the English miles on plains, but fall short of them in hilly grounds. Another opinion is, that the Itinerary miles were not measured by an invariable standard, but in the distant provinces were derived from the common measures of the country. In support of this conjecture a supposed coincidence between 'he computed and measured miles, noticed by Horsley and others, has been adduced; but if this were the case, there would not be so exact a conformity between the miles of France and Italy as appears in the instance before mentioned.

To remove, however, as many causes of error as possible, considerable

XVI. XIX. or XXI. and single numerals might have been omitted, as XX. and XXIII. for XIX, and XXXIII.

• Near Leicester.

Hist. de l'Académie, t. 88, p. 661.

pains have been taken to correct the numbers, by a comparison of all the earliest and most authentic copies of the Itinerary. These are: The Itinerary of Talbot, published in Leland's works. That of Camden. Two copies by Harrison, published first in Hollingshed, and republished by Burton. That of Gale. That of Surira, who collated five copies, four of which he thus designates:-1. Bibliothecæ Regiæ ad D. Laurent. vetustiss. Codex Ovetensis Æra ICCCCXX descriptus. 2. Bibliothecæ Blandiniæ pervetustus codex a CCCC. circiter annis transcriptus. 3. Bibliotheca Neapolitanorum Regum qui post cardinalis de Ursinis fuit anno M.CCCCXXVII. exscriptus. 4. Christophori Longolii exemplar ab H. Stephano. Parisiis editum, anno M.IƆXII.

As the Roman posts and roads were in a great degree connected with, or derived from, the British towns and trackways, we proceed to trace first the course of the British roads which still exist, and to specify the towns whose sites are known, premising that of the ninety-two capital towns of the Britons commemorated by historians, the names of only eighty-eight have been preserved.

The British ways were,

1. The WATLING STREET, or Irish road, in two branches, northern and southern.

2. The IKNIELD STREET, or road of the Iceni, the inhabitants of the eastern coast.

3. The RYKNIELD STREET, leading through the country of the Upper Iceni or Coritani.

4. The ERMYN STREET, leading from the coast of Sussex to the southeast part of Scotland.

5. The AKEMAN STREET, or intermediate road between the Iknield and Ryknield Street.

6. The UPPER SALT-WAY, leading from the salt-mines at Droitwich to the coast of Lincolnshire.

7. The LOWER SALT-WAY, leading from the same mines to the south

eastern coast.

8. A road which appears to have skirted the western coast, as the Ermyn Street did the eastern.

Besides these, there is reason to conjecture from several detached pieces, that another road followed the shores round the island.

WATLING STREET.

The south-eastern branch of the Watling Street proceeded from Richborough on the coast of Kent, to Canterbury; and from thence, nearly in the line of the present turnpike, towards Rochester. It left that city to the right, passed the Medway by a ford, and ran almost straight, through lord Darnley's park, to Southfleet. It bent to the left to avoid the marshes near London, continued along a road now lost to Holwood Hill, the capital of the Rhemi, and then followed the course of the present road to London. Having crossed the Thames, it ran by Edgeware to Verulam; and from thence, with the present great Irish road, through Dunstable and Towcester to Weedon. Hence, instead of bending to the left, with the present turnpike, it proceeded straight by Dovebridge, High Cross, Fazeley, Wall, and Wellington, to Wroxeter. It then passed the Severn, and continued by Rowton, Pen y Pont, and Bala, to Tommen y Mawr, where it