Excavations of 1930

Joined: 9:29 PM - Aug 22, 2006

2:08 AM - Jan 29, 2010 #1

For nearly two centuries, the Society of Antiquaries of Newcastle upon Tyne has produced an annual journal, Archaeologia Aeliana. It is a priceless storehouse of reports on regional history, much of it unavailable anywhere else. All periods are covered, but Hadrian's Wall is the journal's anchor. In the 1930s, Eric Birley launched the first scientific excavations at Vindolanda. And he wrote his reports exclusively for Archaeologia. Though some of his theories have been disproven, and some information has been superseded in the past 80 years, there is much that still rings true. And these first recorded excavations at Vindolanda are still the bedrock on which all of the later work has been set.

Through great generosity by the Society of Antiquaries, We Dig has been given permission to reproduce these excavation reports for the benefit of the community. I will be uploading them as I acquire them, which hopefully won't take terribly long! Please note that these documents are still held under copyright, are owned by the Society of Antiquaries, used here by permission, and not intended to be reproduced or republished.

That said, I hope you will enjoy this, the first report by Professor Birley, from Archaeologia Aeliana, Series IV, Volume VIII (1931), pp. 182-212.
Last edited by SacoHarry on 3:02 PM - Feb 10, 2010, edited 3 times in total.

Joined: 9:29 PM - Aug 22, 2006

2:08 AM - Jan 29, 2010 #2


By Eric Birley.

[Read on 26th November, 1930.]

The following abbreviations are used in this report:
A.A. 1, 2, 3 or 4. Archaeologia Aeliana, series 1-4.
A J. Archaeological Journal.
C.I.L. Corpus Inscriptionum Latinaruon.
C. C.I.L., vol. VII.
C.W. 1 or 2. Transactions of the Cumberland and Westmorland Antiquarian and Archaeological Society, O.S. and N.S.
D. Figure types in Dechelette, Les vases ceramiques ornes de la Gaule romaine.
E.E. Ephemeris epigraphica.
Knorr, 1919. R. Knorr, Topfer and Fabriken verzierter Terra-sigillatades ersten Jahrhunderts.
P.I.R. Prosopographia imperii Romani.
P.-W. Pauly-Wissowa, Realencyclopadie.

In general, a place-name in small capitals refers to an excavation report or museum catalogue that can be readily identified.

In May, 1930, there began at Chesterholm the first of a series of excavations that will last, it is hoped, until the site has been thoroughly explored. It has seemed advisable, therefore, to preface the first interim report on the excavation with an account of the site, and a summary of what was previously known or conjectured about its history.

Dr. Hunter was the first antiquary to visit the place; in 1702 he gave an account of the site in a letter published in the Philosophical Transactions, and described some recent discoveries made to the west of the fort itself, in the bath-house whose smoke-blackened ruins gave the place its name of “the Bowers.” In Hodgson's words, “the Bowers, from the Roman age till within the last century, was the elysium of a colony of fairies; and this ruined bath the kitchen to one of their palaces, of which the soot among the stones was undeniable evidence.”1 Dr. Hunter noted that traces of towers could be made out “on the walls of the fort,” and among the inscriptions published one (no. 3 below) that had “contained the name of the cohors quarta Gallorum, and so proves that this place is Vindolana.”2 Warburton did some digging here in 1717, and found an inscription, now at Durham; it is not clear whether Horsley was ever there--he has nothing to add to Dr. Hunter's account [a curious comment, as Horsley’s work says that he was personally shown the site of the bath house - editor]; the modern history of the place may be said to begin with its acquisition in 1814 by the rev. Anthony Hedley, an early member of this society, who devoted considerable time and energy to the examination of it. Hedley died after a very short illness in January, 1835, and left no record of the work that he had carried out there; for what we know of his researches we depend on the account given by the rev. John Hodgson in part II, volume iii, of his HISTORY OF NORTHUMBERLAND.3 Hodgson was Hedley's close friend for more than twenty years, and his record of the discoveries made at Chesterholm is of especial interest for the light that it throws on the character and ability of that “zealous and warm-hearted antiquary,” the first member of this society to carry out excavations on a Roman site.

In 1818 Hedley's tenant was digging for stone on the east side of the fort, when he came upon a flight of steps leading up from the burn; these steps he removed at once--like most Roman sites in the north, the Bowers had long served as a quarry for stone for every purpose, and Hedley could deplore but not put a stop to the practice. But he followed up the discovery, and excavated the east gate of the fort, to which the flight of steps led.4 In the autumn of 1831, when he had just built or was on the point of building5 the cottage of Chesterholm (from which the fort has since derived its name), he turned his attention to a prominent grassy mound inside the fort, sixty feet northwest of the east gate; there he found an elaborate building--

"several apartments of a sudatory, three of which were built as
usual upon pillars, and the greatest of them measured 21 feet from
east to west, and 13 from north to south, including on that side a
semi-circular recess, on the outside of which the noble altars6 ...
were discovered. . . . The mouth of the furnace to this hypocaust
was about 6 feet in advance to the west of this principal room,
arched, and narrowed in height and width inwards; and strong
marks of fire on its floor, roof, and walls showed that it had been
much used. The pillars to both the rooms were of different shapes
and diameters--some of them portions of square columns, moulded
and fluted on all sides; and some circular, like the banisters of
stairs, as may be seen by the specimens of them in the garden at
Chesterholme; but those of one of the smaller rooms, which was
12 feet square, were shorter than the rest. Adjoining the third
room, and between it and the entrance-room, on the north, was a
cistern 46 inches by 27; and north of it another, 5 feet 5 inches by
4 feet, but not on pillars, both floored and lined with bath cement.
From this range of apartments buildings seemed to have branched
off to the east, west, and south."7

In the absence of a plan, the account presents some difficulties; until the building has been uncovered again it is not necessary to attempt to solve them : but it is clear, at least, that the building was not an early one, since a variety of architectural fragments had been re-used in it : and that there was a street running east and west immediately to the north of it, in addition to the street leading to the east gate, to the south of it. Above all, it is clear that Hedley's excavation involved a careful examination of the structure, and was in a different class to the rather haphazard digging for inscriptions, so characteristic of an earlier, and not unknown in a later, generation of antiquaries.

In the following year and 1833, he cleared the northern part of the west wall of the fort (it was still standing 12 feet high, “but partly bilged out”) and the west gate, in one of the guard-chambers of which he found some 300 coins, “mostly of Constantius and Magnentius, but a few of Constantine II and Constans.” Perhaps it was after this (Hodgson gives no date for it) that he excavated “the towers on both sides of (the north gate), and found them paved at the bottom level with the natural surface of the ground. He also removed the rubbish from a long reach of the north-east portion of the east wall, where the courses remaining are of different thicknesses, and sometimes two run into one; and where the facing stones have not the usual square character of Roman masonry, but, in length, often exceed their height by twice or thrice."8

2 Alexander Gordon has sometimes been credited with this identification, which he puts forward, indeed, as though it were his own; he has clearly taken it from Hunter, to whose account he refers indirectly. Cf. ITINERARIUM SEPTENTRIONALE, p. 78. Bruce ascribes the identification to Horsley.
3 Other descriptions of the site are given by Maclauchlan and Bruce, in addition to the writers mentioned above. Maclauchlan's plan of the fort and its surroundings is valuable; Bruce epitomizes Hodgson, not always accurately.
4 Hodgson, op. cit., p. 196. Cf. also p. 199 below.
5 Hodgson in one place (p. 197) gives 1830, in another (p. 330) 1832, as the date of the erection of this “sweet picture of mosaic work . . . a cottage in the Abbotsford style.” There is evidence that suggests that Chesterholm was completed in 1831.
6 Cf. no. 5 below.
7 Op. cit., p. 196.
8 Op. cit., p. 197. This type of masonry was noted as characteristic of the Theodosian reconstruction at BIRDOSWALD; cf. also below, p. 197.
Last edited by SacoHarry on 1:14 PM - Jan 29, 2010, edited 3 times in total.

Joined: 9:29 PM - Aug 22, 2006

2:09 AM - Jan 29, 2010 #3

Outside the fort, the bath-house was known from the time of Hunter; some seventy-five years later, Wallis records the discovery by some masons of a temple, at the west end of the station, “adorned with Doric pilasters and capitals”; whilst Hodgson mentions two cemeteries, one “in a swampy part of a close to the south-west of the field in which the station stands,” the other “in the fields on the north side of the Causeway.”

Some years after Hedley's death Chesterholm passed into the possession of John Clayton, but no further excavation took place there. Housesteads, Chesters, and the Wall monopolized the attention of the spade, and although no site in the north of England, except perhaps Whitley Castle, presents such promising indications of unrobbed structural remains, the only subsequent discoveries made there up to 1930 were due to agricultural work. Yet Chesterholm has not been ignored, for all that it was not excavated, for its position commands attention.

The fort is a mile due south from Hotbank milecastle, and nearly two miles south-west from Housesteads; it occupies a spur of land, from the north, east, and south sides of which there is a steep fall to the Brackies burn, to the Chainley burn (formed by the junction just below Codley Gate farm of the Brackies and Bradley burns), and to the Doe sike; to the west, the ground rises gently, whilst across the streams on the other three sides there are hills higher than that on which the fort stands. Of these hills, Barcombe shut out all further outlook to the east; Cod Law on the north hides the nearest stretch of the Wall, though Winshields to the north-west, and the crags east of Hotbank to the north-east, are in sight of the fort. To the south there is a good view down the valley of the Chainley burn to the hills across the South Tyne, whilst a stretch of nearly two miles to the west, as far as Seatsides, is in plain view. Some thirty yards north of the fort passes the Stanegate, the ancient road that runs from Carlisle by Nether Denton, Carvoran, and Chesterholm to Chesters or Corbridge or (more probably) to both of these places. It has long been assumed that this road, and the forts on it, were built by Agricola; in recent years the study of Roman pottery has shown that at Carlisle, Nether Denton, Chesters, and Corbridge, there is figured samian, clearly made in South Gaul before the end of the first century, and it has been an easy step to conclude that Chesterholm, lying between undoubted first-century sites on this road, was itself of first-century date.

In his report on the excavations at Corstopitum in 1914, the late R. H. Forster suggested that the end of the first occupation of Newstead and the abandonment of Scotland were followed, somewhere about 90 A.D., by the construction of a chain of forts connected by the Stanegate. The samian found during the excavations at Corstopitum included material later than the time of Agricola himself, and later than any “Agricolan” pottery from Scotland with which Mr. Forster was familiar, yet definitely still of first-century date. Nether Denton presented somewhat similar evidence, and it was only at Carlisle that there was reason to assume an Agricolan, or even perhaps a pre-Agricolan occupation.9 Mr. Forster put forward his view of the date and nature of the Stanegate line “as a theory only, and to some extent as a suggestion of places where excavation might usefully be carried out”;10 the present writer takes this opportunity of acknowledging his indebtedness to that statement of the problem.

Soon after the war Sir George Macdonald put forward the suggestion11 that the Agricolan occupation of Scotland did not come to an end, as Tacitus implies, within a short time of Agricola's recall, but lasted well into the second century--perhaps to the very end of Trajan's principate; this suggestion has since won general acceptance. Yet there is one class of evidence that would seem to conflict so strongly with it as to render it untenable. Already in his 1914 report Mr. Forster could doubt whether the Newstead samian would allow of an occupation lasting much, if at all, beyond 90; and subsequent research, especially since the appearance of Dr. Davies Pryce's admirable report on the samian from Brecon Gaer,12 has tended to emphasize the complete absence from Scotland of the samian characteristic of the time of Trajan.13 This is not the place to say more on a subject that requires elaborate and exhaustive treatment; the writer hopes to collaborate with Dr. Davies Pryce in the near future in a critical examination of the whole of the evidence; for the present it will be enough to say that the Scottish samian provided a strong ground for suspecting that a reorganization of the northern frontier took place in the last years of the principate of Domitian. Now, in recent years, there has been an increasing tendency to put the erection of the Stanegate frontier (though not, of course, the Stanegate itself) into the early years of Hadrian; thus Mr. Collingwood, in his new book, 14 regards the Stanegate and Vallum as parts of one scheme of frontier organization, carried out after the British revolt at the beginning of Hadrian's principate. Such a view of the situation is reasonable, if the continued occupation of Scotland be assumed; but if the samian evidence is to be accepted, it would seem that between c. 90 and 120 there were no Roman garrisons north of Cheviot, and in that case it would be indeed remarkable if no attempt whatever had been made to define the limit of Roman occupation for thirty years, when both Domitian and Trajan devoted considerable attention to frontier organization in Germany and in Raetia. It is perhaps unfortunate that Hadrian's initiation of Roman frontier-works in their most elaborate form should have tended to obscure the real achievements of his immediate predecessors.

In the light of the Scottish evidence it seemed to the writer that excavation on a Stanegate site would show (as Mr. Forster had already conjectured) an occupation beginning at about the same time as the occupation of Scotland appears to stop, and perhaps ending with a disaster, such as is hinted at by Fronto,15 shortly before the reorganization of the frontier by Hadrian; it was in the hopes of establishing or disproving this view that the excavation of Chesterholm was chiefly undertaken.

There are other reasons, however, why this site in particular invites attention. The surface indications have long been known to show that there were considerable buildings to the west of the fort; in 1914 the search for a spring in the western part of the Camp Field revealed a dried-up Roman well (cf. the plan, plate xxxv), and near to it two altars, one of which (no. 6 below) had been set up by the vicani Vindolandesses. Not only does this altar lg give the correct form of the Roman name of the place--Vindolanda--but it testifies to the existence there of a settlement with a corporate existence of its own, such as is only recorded epigraphically elsewhere in Britain at Old an Carlisle. Hitherto, comparatively little attention has been be paid in this country to the settlements outside the Roman forts, yet it is certain that in them will be found the best evidence for the life and culture of the soldiers and their families. The thorough examination of a site such as this is essential for a proper view of the conditions prevailing in the northern military district, and nowhere is there a better prospect for such an examination than at Chesterholm.

Finally, the discovery of a post-Roman tombstone (no. 7 below) suggests the possibility of some continued occupation of the site at a period when most of the forts in the north had been abandoned or destroyed, and the prospect of throwing some light on this dark period in Northumbrian history provides an additional incentive for excavation.

9 So J. P. Bushe-Fox in Archaeologia, LXIV; Haverfield and Atkinson doubted whether the material was so early, but there is no reason to abandon Mr. Bushe-Fox's view.
10 A.A. 3, XII, p. 269.
11 J.R.S., IX, “The Agricolan Occupation of North Britain."
12 In R. E. M. Wheeler, THE ROMAN FORT NEAR BRECON; cf. also T. Davies Pryce, Cornucopia Bowls and Allied Vessels, in THE ANTIQUARIES JOURNAL, X, 4.
13 I see no reason to agree with Mr. Miller's view that the pottery from OLD KILPATRICK includes material of Trajanic date; the bulk of the pieces so dated by him appear to belong to a period subsequent to c. A.D. 130.
15 DE BELLO PARTHICO, Loeb edition, II, p. 22.
Last edited by SacoHarry on 2:45 AM - Jan 29, 2010, edited 1 time in total.

Joined: 9:29 PM - Aug 22, 2006

2:09 AM - Jan 29, 2010 #4


Some forty inscriptions belong to this site, many of them mere fragments. Seven of these may now be considered, for the light that they provide upon the history of the site.16

1. (C. 713; L.S. 260; R.I.B. 1538. At Chesterholin, where until the autumn of 1930 it was built into the north wall of the kitchen passage.)
[Imp. Ca]es. Traia[no I Had]riano[Aug., |
le]g. II[Aug. | A. Platorio Nepote leg.
pr. pr. ]

Hodgson records the discovery of this inscription at Chesterholm, but says nothing of the circumstances of the find. If the attribution is correct, the stone proves Hadrianic building here; but it is perhaps surprising that such an inscription should come from a cohort-fort, where something rather more ornate might be expected. Its resemblance to the milecastle inscriptions is very marked.

2. (A.J. XXXVI, p. 156; E.E. VII, 1050b; R.I.B. 1561. At Chesterholm, where it was re-discovered in 1930 by Mr. James DiacIutyre.)
SV|b sex. calpurnio
AG|ricola leg. aug. pr. pr.

W. Thompson Watkin, who published this inscription in A.J., described it as “found earlier, and now standing, stacked with other remains in the garden.” It is the lower left-hand corner of a large and imposing building-inscription, flanked by an ornamental panel, from whose decoration it is possible to infer that there were originally probably five lines of text. The restoration may be taken as certain, in substance. Calpurnius Agricola was governor of Britain under Marcus and Verus; other records of his governorship have been found at Ribchester (C. 225), Carvoran (C. 758, 773, 774), and Corbridge (E.E. IX, 1381). Whether he was governor of Britain before or after his governorship of one of the Germanies is not known; it has been assumed that he was in Britain first, and that by c. 170 he had one to Germany. Thus, this inscription shows that the fort was rebuilt, at least in part, in 165 or thereabouts.

3. (C. 715; L.S. 262; R.I.B. 1540. Already destroyed in Horsley’s time.)
. . . . . . . . . . coh. IIII]Gallor. | [Alexandriana
de]vota nu | [mi]ni eius por[tam cum tu]rribus |
[a f]undamen[tis restitu]erunt sub | C1(audio)
Xenopho[nte l]eg. Au[g.] pr. [pr.] | curante
[. . . . . . . . . .

Claudius Xenophon was governor in the time of Severus Alexander, c. 222-224 (E.E. VII, 1115; Atkinson, J.R.S. XII), so that the restoration of the cohort's title Alexandriana may be taken as certain. As yet, there is no evidence of rebuilding here under Severus, but it would be unwise to assume that such rebuilding did not take place; at Birdoswald and Risingham, for example, Severan rebuilding was followed by further work in the following decade. For the cohort, cf. no. 5 below.

4. (C. 701; L.S. 257; R.I.B. 1522. Chesters Museum; found in 1838 on the demolition of a cottage at Hardriding, a mile and a half south-west from Chesterholm, whence it had presumably been taken.)
Deo | Cocidio | Decimus | Caerelli | us Victor |
pr. coh. II Ner. | v.s.l.m.

Cohors II Nerviorum appears in the diplomata for 98, 122, 124, and 146; it has left records also at Wallsend (E.E. IX, 1159), Carrawburgh (E.E. III, 103; a vexillation of the cohort), and Whitley Castle, where it was under Caracalla (C. 310) and, no doubt, from then on. The Chesterholm inscription cannot be dated closely, but it should belong to the second century and, to judge by its style, hardly to the end of that century.

5. (C. 704; L.S. 244; R.I.B. 1526. Chesters Museum; found by Anthony Hedley in the building north-west of the east gate of the fort.)
I(ovi) O(ptimo) M(aximo) | ceterisque | diis
immort(alibus) | et gen(io) praetor(i) | Q.
Petronius | Q, f. Fab. Urbicus | praef. coh.
IIII | Gallorum | / / / / / / / / / / | ex Italia | domo
Brixia | votum solvit | pro se | ac suis.

We have more evidence, perhaps, for the history of this cohort than of any other regiment of the garrison of Britain. If Cichorius' view is correct,17 it was in Spain in the first century; by the end of the century it was in garrison at Templebrough, where its presence is attested by tiles, and a number of early inscriptions.18 It appears in the diplomata for 122 and 146, and has kit records also at Castlesteads, Risingham, and Castlehill on the Scottish Wall. At Templebrough, the samian shows that there was an intensive Flavian-Trajanic occupation, followed by a reoccupation of some sort in the second half of the second century. The Castlesteads inscriptions (C. 877, 878) were set up by commanders of the cohort; neither stone can be dated closely. The date of C. 1001, from Risingham, can be arrived at with more certainty. It reads Numinib(us) Augustor(um) coh(ors) IIII Gal(lorum) eq(uitata) fec(it), thus recording building there by the cohort when there were joint Augusti.19 Bruce and other writers have ascribed it to the time of Severus and Caracalla; but the garrison of Risingham at that period was coh. I Vangionum (C. 1003); whilst the simple form and good design of the inscription demand an earlier date. The stone should be dated to the joint principate of Marcus and Verus (rather than to that of Marcus and Commodus); and it may be added to the inscriptions recording reorganization of the northern defences under Calpurnius Agricola or one of his contemporaries. It may be noted here, that both at Risingham and High Rochester we find a cohors quingenaria in garrison in the second century, a cohors milliaria in the third.20 C. 1129, at Castlehill, is a dedication Campestribus et Brittanni(ae) by a prefect of the cohort; the spelling PREF on this stone suggests a date late in the second century. Finally, a number of inscriptions at Chesterholm attest the presence of the cohort here, at least from the time of Severus Alexander, and no doubt from that of Severus--unless it was one of those regiments withdrawn from Scotland by Caracalla; and at the time when the Notitia list item per lineam valli was drawn up, it was still in garrison at Vindolanda. Its movements in Britain may be summarized thus:

Up to c. 120 : Templebrough.
120 to 165 : Castlesteads.
165 to 180 : Risingham.
180 to 196 : Castlehill.
200 onwards : Chesterholm.

Such a scheme is not inconsistent with the existing evidence; it suggests that the fixity of garrisons, well attested from many sites, for years and even centuries on end in one fort, was not always marked in Britain in the second century. As yet, there is hardly sufficient evidence for most regiments, for a detailed scheme of movements in this period to be worth attempting; in general, however, it may be said that what evidence there is shows that no fort on the Wall was garrisoned continuously by the same regiment from the time of Hadrian to the time of Severus; but from the time of Severus on, there is no clear evidence of movements.

The word erased after the name of the cohort was either Alexandriana (cf. no. 3 above) or Antoniniana; in either case it is notable that the commander of the cohort is an Italian.21 Clear cases of Italians commanding auxiliary regiments in the third century are not so common as to be unworthy of special notice, though they are sufficiently numerous to disprove the assertion that from the time of Severus on Italians were debarred from such commands.

I have referred elsewhere22 to the value of this inscription and the similar dedication found with it, for elucidating the meaning of the term praetorium, which is clearly (in a permanent castellum, at all events) the building usually known in England as the “Commandant's House.” Were it only for the discovery of these two inscriptions, Anthony Hedley would deserve the thanks of all Roman historians for his excavation.

6. (A.A. 3, XII, p. 201; R.I.B. 1536. Chesters Museum; found in 1914, 120 yards west of the fort.)
Pro domu | divina et nu | minibus Aug | ustorum
Volc | ano[. . .] sacrum | vicani Vindol | andesses
cur[am] | agente[. . . . . . .] | v(otum) s(olverunt)
l(ibentes) m(erito).

The chief importance of this inscription lies in its spelling of the name of the place; it confirms the reading of Ravennas, Viiidolanda, against that of the Notitia, Vindolana. The name may perhaps mean “Whiteholm” or the like; in the early morning sunlight, such a name would be more appropriate than it seemed to Haverfield. The possibilities of the vicus have been discussed above; it occupies some six acres, and appears to have been so little disturbed by searchers for stone or inscriptions as to offer an exceptionally promising field for investigation.

There are comparatively few inscriptions in Britain mentioning the domus divina, whilst the formula In h(onorem) d(omus) d(ivinae), common on the Rhine, occurs in Britain for certain only at Old Penrith23; the British equivalent seems to be the dedication Numinibus Augusti or Augustorum, which hardly occurs in Germany. The combination of the two forms in this inscription is therefore of some interest. The date off the altar is probably somewhere in the first half of the third century, perhaps in the principate of Severus and his sons.

7. (A.A. 3, XV, p. 29; R.I.B. 1564. Chesters Museum; found east of the fort, in an old field-wall.)

If Haverfield's restoration and identification are correct, this Brigomaglos is the Briomaglos or Briocus who came from Britain to work with Germanus in Gaul, in the early years of the fifth century. In any case, the inscription is very late, and Christian (to judge by the formula); it therefore provides some indication of occupation of the site in the period immediately following the severance of Britain from Rome.

16 In preparing this section I have received invaluable help from Mr. R. G. Collingwood; the numbers prefaced by R.I.B. refer to his forthcoming book on THE ROMAN INSCRIPTIONS OF BRITAIN.
17 S.v. Cohors in P.-W.
18 Cf. May, THE ROMAN FORTS AT TEMPLEBROUGH, where the inscriptions are collected.
19 Haverfield (in ROMAN BRITAIN IN 1914) regarded the plural Augustorum as referring, not to a plurality of living emperors, but to the whole body of emperors, dead and alive. Dedications to the domus divina (for which cf. no. 6 below) do not favour this view; cf. C. 344, from Old Carlisle : I.O.M. PRO SALVTE IMPERATORIS M. ANTONI GORDIANI P(II) P(F,I,ICIS) INVICTI AVG(VSTI) ET SABINIAE FVRIAE TRANQVILI,(AE) CONIVGI EIVS TOTAQVE DOMV DIVIN(A) EORVM. If, however, Haverfield's view is correct, the inscription might very well be as early as the governorship of Lollius Urbicus, and the cohort's period at Castlesteads will in that case be confined to c. 120-140.
20 At High Rochester, coh. I Lingonum was in garrison in the time of Pius (C. 1041), coh. I fida Vardullorum in the third century (C. 1039, etc.). Whether the change implied a larger garrison for forts that were now advanced frontier-posts, and no longer merely on a line of communication behind the lines, is not certain, though it is not unlikely. At present we have hardly sufficient data for determining the relative strength, in the time of Severus, of cohortes milliariae and quingenariae; it might be inferred that by this time all cohorts were of the same strength.
21 The Procurator of Noricum in A.D. 69 was a Petronius Urbicus (cf. P.I.R.); the coincidence of two hardly common names suggests that the procurator was an ancestor of our prefect.
22 C.W. 2, XXX, p. 201.
23 C. 316. Mr. Collingwood tells me that there is also a doubtful example at York. The dated examples of the formula IN H D D on the continent range from the time of Pius (C.I.L. XIII 7458) to that of Diocletian (op. cit. 8019), but it is clearly of considerably earlier origin; the domus divina appears as early as the principate of Tiberius (op. cit. 4635). The Old Carlisle vicus inscription is also a dedication to Vulcan; a good continental parallel is provided by C.I.L. XIII 6454, from Benningen: IN H D D VOLKAN(O) SACRVM VICANI MVRRENSES V. S. L. M. The dedication to Vulcan may be taken to imply that the inhabitants of the settlement were engaged in mining or metallurgy.
Last edited by SacoHarry on 2:29 AM - Jan 29, 2010, edited 2 times in total.

Joined: 9:29 PM - Aug 22, 2006

2:10 AM - Jan 29, 2010 #5


The first season's work at Vindolanda fell into three parts: the uncovering of a number of points on the fort wall, to obtain measurements for the planning of the site; the partial excavation of the north, east, and west gateways (all of which had already been examined by Hedley), and of an area (site C) north of the east gate; and a preliminary examination of the vicus.24 As was bound to be the case in the early stages of the examination of such a site on such a scale, the work was extensive and superficial rather than intensive and exhaustive, and at no point was sufficient excavation carried out to justify a detailed report. The present report aims at presenting the main problems that have arisen in the light of the results so far obtained; for a detailed discussion of structural features or stratification it will be necessary to wait until further excavation has been carried out.

In the fort itself, site C was the first to be dealt with. The east gate that Hedley excavated was considerably farther south than its western counterpart; it seemed possible, therefore, that there might be a second east gate, farther north, and opposite that on the west, and it was hoped that the excavation of such a gateway, where there had been no previous excavation, might provide an epitome of the history of the site. In the event, no gateway was found, but certain features suggested very clearly that there had once been one there. At this point a building was found, close up to the fort wall, and constructed at a high level over the clay rampart-backing, and at least on't occupation layer; its position is reminiscent of the late apse-ended structure at MALTON. Coins and pottery give a satisfactory indication of its date :

Coins.1. A coin of Maximinus Daia wag found beneath the partition-wall between the north and south rooms of the building.
2. A coin of Constans, dateable ta 342-348, was found in the core of the east wall, 16 feet north of the partition-wall.

Pottery.1. (fig. 3, no. 3.) Rim of a Huntcliff-ware cookingpot, heavily burnt. Cf. BIRDOSWALD, no. 20.
2. (fig. 3, no. 4.) Part of a flanged-bowl, as samian form 38, heavily burnt. Cf. BIRDOSWALD, nos. 92, 93. The type occurs in the fourth period only at Birdoswald, at CRAMBECK, and at the Scarborough signal-station; it is thus particularly characteristic of the last thirty years of the fourth century.
[+] Spoiler
Figure 1
The building is thus later than 342-348, whilst the pottery from its floor shows it to have been occupied after 370; it is clearly to be attributed to the Theodosian reconstruction. Its south-east corner is formed by a large block of stone, whilst through-courses of flags occur frequently in its walls; both these features, occurring also in the Theodosian walling at Birdoswald, confirm the evidence of the coins and pottery for dating the rampart-building. Another feature, the lavish use of a distinctive light-brown mortar, is therefore dateable; this mortar was met with again in a large, rough patch in the outer face of the fort-wall at the Southern end of this site; in the patch, flags were used extensively. The fort-wall ran on without a break past the point where a gate-passage was expected, but at this point there is a very pronounced settlement in
[+] Spoiler
Figure 2
the east wall of the rampart-building, which has dipped about two feet, six feet north of the partition-wall : we may suspect that the subsidence has been caused by the insufficient filling of an earlier gap in the rampart--that is to say, that there is a gate-passage below the wall.25 Heavy rain made it impossible to continue the examination of this site, so attention was turned to the north gate, which appeared to have been most completely examined by Anthony Hedley; so that excavation there, whilst it would recover the plan, was not likely to produce such stratification as would suffer from work done in the unfavourable weather then prevailing. The gateway (fig. 2) proved to be of a type unknown in Wall-forts, though met with, for example, at Newstead--a single passage flanked by guardchambers, which in this case project five feet outside the fort-wall. The passage is 12 feet wide and a 21 feet long; the internal dimensions of the guard-chambers average 15 feet by 9 1/2 feet. Two periods were distinguishable. In the first, the gate-passage had a surface of rammed gravel; the masonry of the gateway was all re-used material, with shale (no doubt from the extensive deposits on the north bank of the Brackies burn) used in the place of lime-mortar or clay; I know of no parallel to this feature. In the second period, a roadway of large flags was laid in the passage and on into the interior of the fort, nearly a foot higher than the earlier roadway (cf. plate xxxvii, figs. 1 and 2). In the passage itself the flags had been largely robbed, no doubt during Hedley's examination of the structure26, but farther south they remain in position, and extend for upwards of thirty yards into the fort. The eastern guard-chamber was found to have been completely cleared; that on the west was almost undisturbed. Here an oven had been constructed in the course of the first period (cf. plate xxxviii, figs. 1 and 2, and fig. 2); in the second period a floor partly of flags was laid over the remains of the oven. The only pottery from this floor was a scrap of Huntcliff ware.

The outer corners of the gateway are all dipping very noticeably; that on the north-west27 has collapsed, and has been repaired by the addition of a rough buttress, whose core was full of fallen facing-stones, and bound together with a quantity of the same brown mortar as was noted as characteristic of Theodosian work on site C. At Birdoswald the via principalis was paved, in the same period, with large flags, so that the Theodosian date of the second period in the north gate is confirmed.

24 The work in the vicus took place in the latter part of May and in June; that in the fort, in July. In the earlier work, two men only were employed; in the later, five.
25 The dip in this wall can be seen in plate xxxvi, fig. 2.
26 Cf. p. 185 above.
27 Cf. plate xxxviii, fig. 2.
Last edited by SacoHarry on 1:29 AM - Jan 30, 2010, edited 9 times in total.

Joined: 9:29 PM - Aug 22, 2006

2:10 AM - Jan 29, 2010 #6

During the planning of the fort the south face of the east gate was uncovered. Hodgson's account of the thickness and surviving height of the wall (8 feet and 6 feet respectively) was found to be substantially accurate, and it was found that Hedley had not excavated to the bottom of the passage. There, in a small undisturbed deposit, was a small group of pottery that included the following pieces :
[+] Spoiler
Figure 3
[+] Spoiler
Figure 4
Here, too, the masonry was set in shale; the pottery dropped first after its construction is paralleled by period III pieces at Poltross Burn : the inference is clear that this gateway, and the first period of the north gate, belong to the Constantian reconstruction of c. 300.

It has been shown that the epigraphic evidence attests occupation of the site in the second and third centuries. What of the pre-Constantian fort? On the west side of the north gate there remains the footing, and part of the first course above it, of an earlier fort-wall, two feet farther north than that contemporary with the gateway; and during the planning of the site an earlier south wall was found, at the east two feet, at the west seven, farther south than the existing south wall. Here, then, is clear evidence of an earlier, slightly larger fort, that was very largely destroyed at the time of the Constantian reconstruction. At the north gate, it has been shown, there is evidence for two periods in the existing fort, and at least one earlier : in the fort-wall on site C it seems possible to distinguish three periods. To the north of the clearly recognizable Theodosian patch there is another, extensive patch of large square blocks of ashlar, carefully cut and carefully laid, and quite unlike the Theodosian work. At one point this heavy patch has bulged out considerably, and farther south it has collapsed; here the surviving courses have been propped up with a mass of material, amongst which a quantity of painted wall-plaster was found : and upon them has been put a second patch, built with a considerable batter to prevent it too from falling, and in the characteristic Theodosian style.28 It would seem, then, that the east wall of the last two periods made use of that of the preceding period, whilst the earlier north and south walls were abandoned; and though the evidence is not yet complete, the same would seem to have been the case with the west wall. We have now an indication of Severan, Constantian and Theodosian building (to use the convenient terms for the Wall-periods II, III and IV); but there is also evidence of earlier structures.

At site C, where we have distinguished three periods in the fort-wall, it is noticeable that some of the stones even in the footing-course are re-used material; while that footing runs without a break past the point where it seems probable that there was once a gateway. That gateway should therefore be pre-Severan. Again, the pronounced falling-away of the outer corners of the north gate suggests that there exist, below these corners, the filled-in ends of ditches, such as might be found if there had been an earlier north gate, farther south than the present one, and approached by a causeway flanked by two or more ditches.
[+] Spoiler
Figure 5
It has been shown that at this place the Severan wall is farther north than the Constantian; the north gate that has just been postulated, farther south, must be earlier still. It proved impossible to proceed further with the examination of this gateway; a system of draining will have to be devised, or fine weather secured, before the ditches, and what lower levels there may be, can be excavated.

The preliminary work on the fort was completed by the uncovering, for planning, of the corners of the west gate and of a number of points on the fort-wall. In general type, the west gate (fig. 5) was similar to that on the north, but its guard-chambers are considerably longer in proportion to their breadth (averaging 20 feet by 10 feet); whilst the lay-out is very rough and ready--the gate-passage is 18 inches wider at the outside than at the east end, and the guard-chambers are exceptionally asymmetrical; it would be difficult to find a less accurate piece of planning. The fort-wall proved hardly less irregular; its thickness varied from less than 3 feet to rather less than 8 feet, whilst at one point, south of the west gate (cf. the plan, plate xxxv), an extra face had been added at a high level on the inner side. The masonry varied considerably in character from point to point, and it is plain that the elucidation of the history of the wall alone will involve many problems.29

In the vicus the work done was confined, for the most part, to a site west of the fort, and east of the Roman well discovered in 1914 (site B). First, however, some preliminary trenching was done at the extreme west end of the Camp Field, to determine the westward limit of occupation; the results obtained must be supplemented by further digging before their significance can be determined.

“The station was plentifully supplied with water,” Hodgson writes30, “by a channel cut in large stones from a copious spring, about a furlong to the west. Mr. Hedley, in 1832, found several roods of this gutter stone lying quite perfect, and near the surface.” Parts of this water-channel were still visible in May, 1930, when excavation began, and attention was soon turned to it. At one point (a on the plan, fig. 6), the channel-stone had been removed, and it was noted that there was forced soil below where it had been. A trial-hole sunk here reached the undisturbed subsoil over four feet below the present surface, in the bottom of a V-shaped ditch, running approximately north and south; in the ditch, besides a considerable peaty accumulation, was some first-century pottery, including part of a South Gaulish bowl, form 37 (F.S. 1 below).

Almost at the start of the excavation, evidence had come to light of the Agricolan occupation that had long been postulated. Further trial-holes were made, as a result of which it was possible to trace a stretch of some 60 feet of this ditch (whose course is plotted approximately on fig. 6);
[+] Spoiler
Figure 6
and from each hole came more first-century pottery, including most of the pieces shown in fig. 7. The surviving dimensions of the ditch varied considerably, as the ground had been levelled after it went out of use; its sides sloped at an angle of rather more than 45 degrees, and its greatest surviving depth was about 2 feet 6 inches. The figured samian found in it is discussed below, because of its special value for dating the early occupation of the site; the other pottery, which includes equally early material, is being held over until the examination of this part of site B has been completed : it includes a good early “rustic” cooking-pot, carinated bowls with reeded rims and well-marked foot-stands, flat-rimmed mortaria, Belgic ware, and other distinctive “Agricolan” types.


At the southern end of site B, the north-west angle of a large stone building was uncovered; its wall still stands 2 feet high, and there is more than 2 feet of forced soil below its foundations. Here, as at a point 20 feet farther north,32 traces were found of wooden structures--postholes, with parts of the posts still remaining in them, and sleeper-trenches; at the latter point there was a wooden building only just outside the early ditch, the heel alone of which survived there. Detailed discussion of the finds there is deferred until the examination of this site has been completed.


Mr. Percy Hedley has supplied the accompanying notes on the five coins found during the excavations. Nos. 3, 4 and 5 come from site C, 3 and 5 being stratified (cf. p. 196 above); no. 1 was found outside the north gate, east of the east guard-chamber : it is much worn; no. 2 was resting immediately on the flagged roadway, 6 feet south of the north gate.

28 Cf. plate xxxvi, fig. 2.
29 The dimensions of the fort, from outer face to outer face of its wall, come to 5o8 by 306 feet, giving an area of upwards of 3 1/2 acres; thus it is rather larger than Great Chesters-Aesica, though it has been called Little Chesters to distinguish it from that fort.
30 Op. cit., p. 195.
31 (removed to “Figured Samian” page --- still to be uploaded)
32 b on the plan, fig. 6.
Last edited by SacoHarry on 12:21 AM - Feb 02, 2010, edited 4 times in total.

Joined: 9:29 PM - Aug 22, 2006

2:11 AM - Jan 29, 2010 #7


It will be seen that although the first season's work was, necessarily, confined to a preliminary sampling of the possibilities of the site, sufficient results were obtained to amplify what was already known about its history in some detail. The “Agricolan” occupation is now established, though its initial date is not yet certain, and structural remains of the fort of this period have yet to be discovered; the early ditch in the vicus, undoubtedly connected with the first occupation, looks to belong rather to the annexe than to the fort itself : and the discovery of two pieces of South Gaulish figured samian, as chance finds, at the north gate and on site C, suggests that the Agricolan fort may have occupied part, at least, of the area on which the present fort stands. Whether or not the first occupation began as early as the governorship of Agricola himself, it certainly continued longer, to judge by the samian, than the Agricolan Occupation of Scotland; among the South Gaulish ware, no. 9 in particular, with parallels at Vindonissa, and from Central Gaul no. 11--a type particularly (and, it would seem, exclusively) characteristic of the first quarter of the second century, are without Scottish parallels.

Whether the site was re-occupied under Hadrian, it is not yet clear. The Hadrianic inscription, if it belongs to the site, would be decisive; but as yet no pottery has been found that could well be dated to the period 120-150. When so small an amount of digging has been done, however, and that confined for the most part to higher levels, it would be unwise to draw any conclusions from negative evidence. In any case, soon after 160 the site was in occupation; it was involved in the general destruction that overtook all the sites on the north in 196, when Albinus and his army were warring against Severus in Gaul; and when it was rebuilt, the east gate was probably abolished. By 222-224, the fort had been thoroughly reconditioned, and from now on it was held by coh. IIII Gallorum-a cohors equitata, whose mounted men would be specially suited for the patrol and convoy work that must have bulked largely in the work of this fort; Vindolanda was reckoned, indeed, as one of the stations per lineam valli, but its position--with no clear evidence of direct communication with the Wall--is rather that of a line of communication post.

Towards the end of the third century there followed another destruction, so complete that the Constantian repairers could raze the old north and south walls, thus reducing the length of the fort, and giving it its present form : the plan, plate xxxv, shows the outline of this fort, based on the 1930 excavations (as yet, no examination of the south gate has been attempted, and therefore a wide gap is left on the plan, where the surface indications show it to be). After the Picts' war, the fort was rebuilt by Theodosius (or rather by Dulcitius, who as dux had special charge of the northern command); now large, rough, but solid patches were made in the fort-wall, and one at least of those barrack-buildings was erected, that seem to be the nearest British equivalent to the late-Roman system of which Altrip is a typical example.33 To judge by the burnt pottery from this building, this period ended (as it ended at Birdoswald) in destruction by fire; whether this was in the was with the Picts destruction that occurred at the beginning of Maximus's command, or after his withdrawal of troops to the continent, it is not yet possible to say. So far, also, there is no trace of a later, non-military occupation, such as might be associated with Brigomaglos.

In the vicus, besides the early ditch, we have traces of early timber-buildings, and also of a stone-building, the date of which is probably (to judge by such pottery evidence as has been obtained) soon after A.D. 200. The insertion of buildings inside forts in the closing years of the Roman occupation, suggests that life outside their walls had become too insecure. There is already some ground for suspecting that on the limes, at least, the Roman soldier lived with his family inside the fort, not outside it; subsequent work in the vicus must show whether such was the case at Vindolanda.

The excavations in the fort in July formed part of the programme of the Durham University Excavation Committee, and were carried out at the committee’s cost. Our member, Mr. John Charlton, B.A., of Armstrong College, shared in the supervision of the work for the whole period, and undertook the planning; the plans and sections that he has drawn speak for the value of his co-operation. During a lesser part of the time, Mr. James MacIntyre and Mr. C. E. Stevens also assisted in the supervision, besides taking an active part in excavating. Five students, from the universities of Oxford, Cambridge and Durham, stayed for varying periods and gave general assistance. Mr. F. G. Simpson and Mr. I. A. Richmond found time to make several visits to the site, in spite of their preoccupation with the excavations at Birdoswald; to Mr. Richmond's experience in disentangling different periods in the walls of Rome I am deeply indebted for the views on the fort-wall at site C that have been expressed above. I have also to thank Mr. Collingwood for his assistance in the preparation of the section on the inscriptions; Mr. J. A. Stanfield for his drawings of the figured samian; Mr. Percy Hedley for his detailed report on the coins; Dr. Th. Eckinger, of Brugg, for much assistance in the examination of the material in the Vindonissa Museum; and Mr. John Gibson, for allowing me to reproduce, as plate xxxiv and plate xxxvi, fig. 1, two of his exceptionally good photographs of the site of Vindolanda. The local men engaged for the excavation did very well, for all that they had no experience of similar work, Thomas Batey in particular showing a marked flair for dealing with stratification; and Mr. Thomas Hepple's experience and skill were invaluable.


In conclusion, it will perhaps be as well to state the principles on which it is proposed to construct future reports. As far as the pottery is concerned, pieces of value for dating will be published as they occur; other pieces only in so far as they possess unusual features. Small finds in general will be published intermittently, as sufficient material is collected for separate treatment of different classes of object; but it is hoped to publish such examples of signed figured samian as may be found, regularly, because of the special value of such pieces for comparative purposes. Stratified coins, and coin-hoards, will be recorded in full; other coins not necessarily so fully. It is hoped to keep all the material, published or unpublished, readily accessible for study at Chesterholm.

33 Cf. Sprater, DIE PFALZ UNTER DEN ROEMERN, Speier am Rhein, 1929, pp. 38 et seqq.
Last edited by SacoHarry on 2:37 AM - Jan 29, 2010, edited 1 time in total.

Joined: 9:29 PM - Aug 22, 2006

2:38 AM - Jan 29, 2010 #8

Plates from report:
[+] Spoiler
[+] Spoiler
Plate XXXV
[+] Spoiler
Plate XXXVI, Fig. 1
[+] Spoiler
Plate XXXVI, Fig. 2
[+] Spoiler
Plate XXXVII, Fig. 1
[+] Spoiler
Plate XXXVII, Fig. 2
[+] Spoiler
Plate XXXVIII, Fig. 1
[+] Spoiler
Plate XXXVIII, Fig. 2
Last edited by SacoHarry on 2:43 AM - Jan 29, 2010, edited 1 time in total.