Page images
PDF
EPUB
[merged small][merged small][merged small][merged small][merged small][merged small][ocr errors][merged small][ocr errors][merged small][merged small]

* LIMITS OF TRENCH. 11

= ADVENTITIOUS 5016
BLANK= RAMPART.

E CHARCOAL.
= SUBS016.

[ocr errors][ocr errors]

XX

20

35
1

60

[ocr errors]

FEET.

DATUM LINE ABOUT 255' ABOVE SEA-LEVEL.

RAMPART SECTION AT A-B.

A SUPPOSED ROMAN “CAMP" NEAR

HARROGATE.

By FRANCIS VILLY.

for a

This earthwork, situated about half on Grange Farm and half on Cow Dyke Farm, and about it miles north of Harrogate Station, is noted in various topographical books (e.g. Speight's Nidderdale, p. 315, and Grainge's History of Harrogate, p. 363), and is usually regarded as of Roman origin. It is so classed in the preliminary schedule of the Earthworks Committee of the Yorkshire Archæological Society. The accompanying plan shows that, so far as external appearances go, such a suggestion is highly reasonable ; but the slight investigation to be recounted almost proves it to be mistaken.

To describe the site first, measured from the inner edge of the fosse (so far as the position of that line can be laid down), it covers just about three acres—a very usual size

“cohort-fort,” i.e. one calculated to hold a garrison consisting of an auxiliary cohort of about 500 men, and its almost equilateral form is also usual in forts of this size. The skew shape is strongly in favour of its being Roman work, nor is the slight irregularity of the plan necessarily a contra-indication. The width of the vallum and the regular and widely sweeping curve of the corners are not often to be met with in earthworks of other origin. Situated at an average of about 275 feet above sea-level, it is planted astride of a spur of high ground sloping from west to east, the western vallum being overlooked at rather unpleasantly close range, especially at the north-west corner, the summit of the hill being about 100 feet higher at one-third of a mile away. This military weakness is the only external sign pointing against the work being Roman, and it is far from conclusive by itself. The other three sides face downhill, and are strong. The Oak beck runs approximately north and south at the foot of the slope to the east, nearly 300 yards away. No signs of gateways can be traced.

The whole extent is marked on the 1847 six-inch Ordnance Survey map, the latest one showing only the south-eastern

[merged small][ocr errors]

quarter ; however it is still plainly traceable for nearly threequarters of its extent, the south-western quarter and some other parts being quite obliterated. The plan has been drawn mostly from external measurements (different conventions being used to show to what extent the line is only an assumed one) ; and those parts now invisible are inserted mainly on the evidence of the old Ordnance Survey map. It will be understood, therefore, that accuracy in detail is not guaranteed.

So far as can be ascertained, no relics have ever been found in the vicinity, nor is there any tradition regarding the “camp”-in fact, its very existence seemed to be almost, if not quite, forgotten locally when the investigation was begun. The point which had been noticed was the presence of a curious pit in the line of the western vallum and fosse. So far as can be judged, this is later than the earthwork, for the vallum is broadened out in its neighbourhood as if by the deposition of upcast. It may possibly have once been a quarry, and it is not easy to connect it directly with any sort of earthwork.

Such was the evidence in the possession of archæologists before October, 1911, when I undertook to cut a few trenches (confined to the south-eastern quarter), in order, if possible, to prove or disprove the Roman origin of the place.

I figure the plan of a section cut through the fosse and most of the vallum, just clear of the curve of the south-eastern corner to the north. The soils, whether previously untouched or not, are of a very mixed character, consisting mostly of small and brittle flakes of stone mixed with more or less sand of various colours. This composes the vallum, and is also to be found at each side of the fosse, so that the former seems pretty clearly to have been derived from the latter. The material is quite unsuitable for a rampart, for it would crumble very readily under foot-wear or weathering. A distinct layer of small scattered pieces of charcoal ran, as marked, through the vallum in this section. However it got there, its presence seems to prove that the layers below it have never been disturbed since the bank was first thrown up. The clear berm of five feet in width, a most unusual feature, should be noted. The fosse seemed to have been cut to a point (as shown tentatively in the plan), and this is the typical shape for a Roman fosse. However, there was some doubt as to the exact line in the lower two feet, where charcoal (present frequently in the upper layers of filling) was not observed. The line possibly may have followed a curve, as in the third cut, instead of going to a point.

The next trench was dug at a similar distance from the south-eastern corner towards the west. The vallum here was cut into for thirteen feet, and consisted entirely of sand. Charcoal was again present in scattered particles, but not in any particular layer. The berm was as before. The fosse was only opened through its inner third.

The third cut was radial through the corner, and concerned the fosse only, which here was of the same width as in the first cut, but only three feet deep from the subsoil level (four feet to the sod), and filled with black earth containing many small pieces of charcoal and a considerable number of potsherds right down to the rounded bottom. It is not necessary to describe these in detail, for all date from later mediæval times (about 1500–1700 A.D.), carrying bright green, light green and almost transparent, rich chocolate or yellow to light brown glaze. One bowl and a good many fragments of pipe stems were also mixed in.

It was felt that these features, combined with an absence of any earlier remains, almost, if not quite, negatived the Roman nature of the earthwork, and, with one exception, further work was abandoned in consequence. This exception was a trench sixty-two feet long, parallel to and about six feet east of the wall dividing the southern half of the “camp” in a roughly north to south direction. A little charcoal was found here also, and perhaps a rough paving existed some 30 feet outside the line of the outer edge of the fosse ; but there was no fosse, no sign of construction within its line, and nothing else to indicate that the ground had ever been occupied. The surface soil was only some four inches thick, and was sharply differentiated from the bright yellow sand beneath — most unusual features in a site where anything like a serious occupation has ever had place. We may sum up the evidence obtained by excavation thus :

1. Rampart.---The rampart of a Roman fort is usually constructed in the case of an earthwork) of clay, sods, or other carefully selected material, very frequently on a stone foundation. This rampart appears never to have been disturbed, yet it presents nothing resembling these features.

2. A clear berm of five feet is very unusual in a Roman earthwork.

[ocr errors]

IT

ar: Bewerez ispany traceable for nearly three

6 serent the south-western quarter and some other use iterated. Tie par has been drawn mostly Sr Himenis erect conventions being used

SIT: 32: Elate Des Casar assumed one) ; rose 22 orei mainly on the eviCerer Seray It will be understood, Seir 2 KEY = Sesat garat teed.

2 Tiai Does tave ever been 26 ST regarding the

Isa sosed to be almost, : nein ICETTE znaczation was begun. ***.17 Baldostas e resence of a curious

CEWET T-Diaze. So far as can
S KRT65 for the vallum

cats the deposition SIT DY a quam. and it is T.R srary PILT SO di cart work.

27.5C ii archeologists VR : mir: Cui a ft trenches SYSTI

che, if possible, ལས། ལ་ ་

15.00 ave.

11 ar tre fusse and TTC Ei sictheastern \ I I. Qr untouched

Co mostly of sot Ti. Ir ne less sand Sims Tu and is also to

o fiches setms ལམཝ པ ལ ་ རྫ r atter. The 11 & 1112 r.cd crumble

Asunct laver

arted through *** fit , its presence on si f DET DETET been dis

The citar berm UTINlink tiu sau be noted. 11 1 I PUT AS Shawn tenta

**N11 Sa io a Roman

Max* tact line in 4. Yoodleed. TT: tepently in the

ne possibly

[ocr errors]

11NI:

[ocr errors]
[ocr errors]
[ocr errors]
« PreviousContinue »