FOSSE LANE 1996: SPECIALIST REPORTS
THE ROMAN SETTLEMENT AT FOSSE LANE,
SHEPTON MALLET:
THE TESCO EXCAVATION, 1996–7
Peter Ellis and Peter Leach
with contributions by
Umberto Albarella, Lynne Bevan, Brenda Dickinson, Andy Hammon, Kay Hartley,
Birgitte Hoffmann, Julie Jones, Joanna Mills, Stephen Minnitt, Donald Mackreth,
Stephanie Pinter-Bellows, Fiona Roe, and Roger Tomlin
SOMERSET ARCHAEOLOGY AND NATURAL HISTORY VOL. 155
© Somerset Archaeological and Natural History Society and the authors, Somerset Heritage Centre,
Brunel Way, Norton Fitzwarren, Taunton TA2 6SF, 2011
1
SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011
Contents
Flint Lynne Bevan .................................................................................................................................
The coins Stephen Minnitt .....................................................................................................................
Roman small finds Lynne Bevan ..........................................................................................................
Copper alloy objects .................................................................................................................
Silver object Roger Tomlin .........................................................................................................
Iron objects ..................................................................................................................................
Lead objects .................................................................................................................................
Worked bone ................................................................................................................................
Shale ..............................................................................................................................................
Wall plaster ....................................................................................................................................
The brooches Donald Mackreth ..............................................................................................................
The glass Birgitta Hoffmann .................................................................................................................
The Roman pottery Annette Hancocks ...............................................................................................
Mortarium Kay Hartley ..............................................................................................................
Samian: decorated wares Joanna Mills .......................................................................................
Potters’ stamps Brenda Dickinson .............................................................................................
Graffiti Roger Tomlin ...................................................................................................................
The worked stone Fiona Roe ...............................................................................................................
The charred plant remains Julie Jones ..................................................................................................
The animal bone Umberto Albarella and Andy Hammon ......................................................................
The human skeletons Stephanie Pinter-Bellows ....................................................................................
W3
W3
W8
W8
W10
W10
W11
W12
W12
W12
W12
W14
W18
W24
W25
W30
W31
W33
W35
W37
W44
List of figures
Fig. 29
Fig. 30
Fig. 31
Fig. 32
Fig. 33
Fig. 34
Fig. 35
Fig. 36
Fig. 37
Fig. 38
Fig. 39
Copper alloy and shale objects; scale as shown .................................................................
Vessel glass; scale as shown ............................................................................................
Romano-British pottery; scale as shown ..............................................................................
Decorated samian; scale as shown ..................................................................................
Samian stamps; scale 2:1 ................................................................................................
Animal bone: relative importance of the main domesticates ............................................
Animal bone: relative abundance of skeletal elements ......................................................
Animal bone: post-cranial proportions of butchery .............................................................
Animal bone: cattle survivorship curve (from mandibles with 2+ teeth) for earlier
Roman periods ..................................................................................................................
Animal bone: inter-site comparisons of cattle survivorship curves (from mandibles
with 2+ teeth) for earlier Roman periods .........................................................................
Animal bone: inter-site comparison of cattle M3 for earlier Roman periods ..................
W9
W15
W25
W27
W31
W41
W42
W42
W43
W43
W44
List of tables
Table 4
Table 5
Table 6
Table 7
Table 8
Coin list .................................................................................................................................
Romano-British pottery: fabric, sources and quantities in assemblage .............................
Romano-British pottery: summary of fabrics by period ...................................................
Charred plant macrofossils occurrence ............................................................................
Animal bone: numbers of animal bone and teeth ISP including hand and sieve
collection) ..........................................................................................................................
Animal bone numbers (NISP) and % of the three major domesticates ..........................
Animal bone: comparison of cattle and sheep measurements from selected RomanoBritish sites ........................................................................................................................
Human bone: condition and degree of completeness of skeletons ...................................
Human bone: demography for Shepton Mallet ...............................................................
Human bone: stature, means and ranges for Fosse Lane .................................................
W4
W20
W22
W38
References ...........................................................................................................................................
W48
Table 9
Table 10
Table 11
Table 12
Table 13
W2
W40
W40
W44
W45
W45
W46
FOSSE LANE 1996: SPECIALIST REPORTS
SPECIALIST REPORTS
FLINT Lynne Bevan
THE COINS Stephen Minnitt
Thirty eight items of humanly worked flint were
recovered comprising three cores, two retouched
blades, eight scrapers, and 25 flakes; none are
illustrated. The raw material was of a generally good
quality with the brown, compacted cortex
characteristic of flint from secondary deposits, such
as river gravels. It was beige and light to medium grey
in colour with a high incidence of white recortication.
Pebble flint is often of unpredictable quality, and a
number of hinge fractures, common when working
poor quality flint, were apparent among the flakes and
cores. The three cores in the collection had been
worked beyond the point of apparent usefulness which
is another indication that good-quality flint was at a
premium. The core, blade, scraper and flint find spots
gave no indication of any activity focus and were
evenly spread between areas A and B.
While none of the flints were chronologically
diagnostic, a generally later prehistoric date during
the later Neolithic to early Bronze Age seems most
probable, based upon the broad, squat shape of most
of the flakes. In contrast to the material found to the
south (Bevan 2001a, 105), there was no clearly
Mesolithic flintwork. The presence of eight scrapers
in the collection is suggestive of habitation foci within
the vicinity of the site, but the generally low incidence
of flint tools, cores and flakes is not indicative of
occupation of any longevity. Instead, this small
collection appears to represent a low density and
episodic usage of the landscape throughout later
prehistory, supporting the conclusions drawn from the
1990 collection (ibid, 106).
A total of 298 coins were found of which 296 were
Roman (Table 4). Forty-seven derive from excavated
contexts, the remainder are unstratified. Coins marked
* were identified by Simon Esmonde-Cleary. In view
of the detailed analysis of coins found during earlier
work at Fosse Lane (Esmonde-Cleary 2001, 211) little
comment will be made on the 1996 assemblage.
Suffice it to say that the general pattern of coins found
in 1996 is closely comparable to the earlier finds and
continues the trend of a predominance of coins of 4thcentury date. The occurrence of the Republican
denarius should not be considered as particularly
significant as such pieces were current into the 3rd
century. Unless stated otherwise in the context column
the coins are unstratified.
The assemblage includes an abnormally high
proportion of unidentifiable coins. In spite of careful
and individual cleaning of the coins by Mark Davis,
Keeper-Conservation, Somerset County Museums
Service, surface corrosion (patina), and therefore
detail, failed to adhere to the coins. Whilst in part
this may be due to soil conditions a significant factor
was probably the post-excavation treatment of the
coins. Each was individually wrapped in tissue,
perhaps while it was still damp, and placed in a
polythene bag shortly after excavation. This and a
delay of two years prior to cleaning may well have
had an adverse effect on the stability of the surface
corrosion.
W3
3
SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011
TABLE 4: COIN LIST
Reign
Reverse
Date
D. Silanus, denarius
Victory in biga
91BC
Probably copy of Claudian as, though of unusual style
Vespasian, dup.
illegible
69-79
Hadrian, sest.
Pont Max Tr Pot Cos SC 117
Fort Red
Hadrian, sest.
Pont Max Tr Pot Cos III SC119-122
Hadrian?, sest.
117-138?
Antoninus Pius, dup.
Libertas
138-161
Faustina I, sest.
Aeternitas
141+
Faustina I, sest.
Aeternitas
141+
Marcus Aurelius, sest.
illegible
161-180
Lucilla, denarius
Iuoni Lucinae
164-169
C1-C2 as or dup.
C1-C2 as or dup.
Gallienus
Gallienus
Gallienus
Gallienus
Gallienus
Claudius II
Claudius I I
Claudius II, divo
Tacitus
Victorinus
Victorinus
Victorinus
Victorinus
Victorinus
Tetricus I
Tetricus I
Tetricus I
Tetricus I
Tetricus I
Tetricus I
Tetricus I
Tetricus I
Tetricus II
Tetricus II
Tetricus II
Radiate
Radiate
Radiate
Radiate
Radiate
Radiate
Radiate
Radiate
Radiate
Claudius II, copy
Claudius II, copy
Claudius II, copy
Claudius II, copy
Victorinus copy
Victorinus copy
Victorinus copy
W4
illegible
illegible
Conservat Pietat
Dianae Cons Aug
Iovi Propugnatori
Aetern Aug
Pax Aug
illegible
illegible
Consecratio, altar
Salus Aug
Aequitas Aug
Pietas Aug
Pax Aug
illegible
illegible
Fides Militum
Hilaritas Augg
Laetitia Augg
Pietas Augg
Salus Aug
Spes Aug
Salus?
illegible
Salus Aug
Pax Aug
illegible
Comes Aug
Aequitas
Consecratio, altar
Consecratio, altar
Consecratio, altar
Consecratio, altar
Invictus
Pax
illegible
260-268
260-268
260-268
260-268
260-268
268-270
268-270
c.270
265-266
268-270
268-270
268-270
268-270
268-270
270-274
270-273
270-274
270-273
270-273
270-274
270-274
270-274
270-274
270-274
270-273
260-280
260-280
260-280
260-280
260-280
260-280
260-280
260-280
260-280
270-290
270-290
270-290
270-290
270-290
270-290
270-290
Context
Ref
S.F. No.
Crawford 337/3
19
512
43
192
F 101
RIC 541a
RIC 561a
as RIC 950
3010 structure 6 RIC 770
SW room
2062
2041
2006, F224
RIC 171a
as RIC 176
RIC 214
RIC 465a
RIC 575
as RIC 259
RIC 57
RIC 40
as RIC 58
as RIC 118
as RIC 79
RIC 87/88
RIC 108
as RIC 121
as RIC 130
2098
RIC 266
RIC 247/248
2009
2034
copy as RIC 112
157
182
382
89
231
365
513
366
368
438
325*
179
591
413
102
234
261
311*
169
271
266
226
270
354
87
454
80
273
240*
582
424*
391
351
475
28
185
65
215
235
430
431
448
580
170
284*
285
295*
10
176
376
FOSSE LANE 1996: SPECIALIST REPORTS
Reign
Reverse
Date
Context
Ref
S.F. No.
Tetricus I copy
Tetricus I copy
Barbarous radiate
Barbarous radiate
Barbarous radiate
Barbarous radiate
Barbarous radiate
Barbarous radiate
Barbarous radiate
Barbarous radiate
Barbarous radiate
Barbarous radiate
Barbarous radiate
Barbarous radiate
Barbarous radiate
Barbarous radiate
Barbarous radiate
Carausius
Carausius
Maximian, follis
but obv legend:
MAXIMIANVS NOB C
Licinius
Constantine I
Constantine I
Constantine I
Constantine I
Crispus
Crispus
Constantine II
Constantine I
Constantine II
Constantine II
Constantine II
House of Constantine
House of Constantine
Urbs Roma
Urbs Roma
Urbs Roma
Urbs Roma
Urbs Roma copy
Urbs Roma copy
Constantinopolis
Constantinopolis
Constantinopolis
Constantinopolis
Constaninopolis
Constantinopolis
Constantinopolis copy
Constantinopolis copy
Constantinopolis copy
Constantinopolis copy
Constantinopolis copy
Constantine II
Constantine II
Constans
Constantius II
Constantius II
Laetitia
Pax
Pax
Victory
Salus
Salus
Pax
Pietas?
270-290
270-273
270-290
270-290
270-290
270-290
270-290
270-290
270-290
270-290
270-290
270-290
270-290
270-290
270-290
270-290
270-290
287-293
287-293
305-307
2020
copy as RIC 86
287*
61
105
243*
254*
370
318*
573
18
37
183
256*
265*
337
472
546
590
228
455
299*
Provid Aug
illegible
Genio Populi Romani
Genio Pop Rom
310-312
Victoriae Laetae Princ Perp 319
Victoriae Laetae Princ Perp 319
Beata Tranquillitas
321
Beata Tranquillitas
321
Beata Tranquillitas
321
Beata Tranquillitas
321-323
Beata Tranqlitas
323-324
Providentiae Caes
324-326
Gloria Exercitus 2 stds
330-335
Gloria Exercitus 2 stds
330-337
Gloria Exercitus 2 stds
330-335
Gloria Exercitus 2 stds
330-335
Gloria Exercitus 2 stds
330-335
Wolf and twins
330-335
Wolf and twins
330-335
wolf and twins
330-335
Wolf and twins
330-335
Wolf and twins
330-348
Wolf and twins
330-348
Victory on prow
330-335
Victory on prow
330-335
Victory on prow
330-335
Victory on prow
330-335
Victory on prow
330-335
Victory on prow
330-335
Victory on prow
330-348
Victory on prow
330-348
Victory on prow
330-348
Victory on prow
330-348
Victory on prow
330-348
Gloria Exercitus 1 std
335-337
Gloria Exercitus 1 std
335-337
Gloria Exercitus 1 std
337-341
Gloria Exercitus 1 std
335-337
Gloria Exercitus 1 std
337-341
2006, F224
2005, F223
2068
2004, F213
2009, F224
RIC 356
2029
2020
2060
2058
2063
2005, F223
2020, F223
1036, F122
2074
2060
RIC VI Trier 642a
RIC VI London,209c
RIC VII London, 209
RIC VII Trier, 213
RIC VII Trier, 303
RIC VII Trier, 305
RIC VII Trier, 308
180
200
46
162
289*
160
76
RIC VII London, 287 199
138
as LRBC I, 49
248
as LRBC I, 49
131
149
167
as LRBC I, 48
327*
LRBC I, 51
433
LRBC I, 51
577
as LRBC I, 51
25
as LRBC I, 51
322
copy as LRBC I, 51 314*
copy as LRBC I, 51 198
LRBC I, 71
216
LRBC I, 86
320*
142
LRBC I, 185
377
LRBCI, 191
410
LRBC I,201
541
copy as LRBC I, 52 252*
copy as LRBC I, 52 286*
copy as LRBC I, 52 401*
copy as LRBC I, 52 568
copy as LRBC I, 52 165
LRBC I, 93
392
383*
LRBC I, 131
321*
LRBC I, 94
117
LRBC I, 252
439
W5
SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011
Reign
Theodora
Theodora
Helena
House of Constantine
House of Constantine
House of Constantine
House of Constantine
House of Constantine
Constans, copy
H. of Constantine copy
H of Constantine copy
H. of Constantine copy
H of Constantine copy
515
SW room
Constantius II
Constantius II
Constantius II
Constans
Constans
Constans
Constans
Constans
Constans
House of Constantine
House of Constantine
House of Constantine
House of Constantine
House of Constantine
House of Constantine
Constans
Reverse
Pietas Romana
Pietas Romana
Pax Publica
Gloria Exercitus 1 std
Gloria Exercitus 1 std
Gloria Exercitus 1 std
Gloria Exercitus 1 std
Gloria Exercitus 1 std
Gloria Exercitus 1 std
Gloria Exercitus 1 std
Gloria Exercitus 1 std
Gloria Exercitus 1 std
Gloria Exercitus 1 std
Date
337-341
337-341
337-341
335-341
335-341
335-341
335-341
335-341
335-348
335-348
335-348
335-348
335-348
Context
3014
Ref
as LRBC I, 105
as LRBC I,120
LRBC I, 112
as LRBC I, 87
as LRBC I, 139
As LRBC I, 107
as LRBC I, 100
copy as LRBC I, 87
2001
S.F. No.
272
55
171
511
94
219
594
589
59
437
114
474
3010 structure 6
Victoriaeddauggqnn
Victoriaeddauggqnn
Victoriaeddauggqnn
Victoriaeddauggqnn
Victoriaeddauggqnn
Victoriaeddauggqnn
Victoriaeddauggqnn
Victoriaeddauggqnn
Victoriaeddauggqnn
Victoriaeddauggqnn
Victoriaeddauggqnn
Victoriaeddauggqnn
Victoriaeddauggqnn
Victoriaeddauggqnn
Victoriaeddauggqnn
Fel Temp Reparatio,
phoenix
House of Constantine
Fel Temp Reparatio,
phoenix
Magnentius/Decentius fragment, Felicitas Reipublice
Magnentius
Gloria Romanorum
Magnentius
Victoriae dd aug et cae
Magnentius?
Constantius II
Fel Temp Reparatio (FH)
Constantius II
Fel Temp Reparatio (FH)
Falling horseman copy
Falling horseman copy
Falling horseman copy
Falling horseman copy
Falling horseman copy
Falling horseman copy
Falling horseman copy
Falling horseman copy
Falling horseman copy
Falling horseman copy
350
350
351-352
350-353
353-355
350-355
350-360
350-360
350-360
350-360
350-360
350-360
350-360
350-360
350-360
350-360
Falling horseman copy
350-360
3013 str 6
SE room
Falling horseman copy
Falling horseman copy overstruck on Gloria Exercitus
two standards
Falling horseman copy overstruck on Gloria Exercitus
Valentinian I
Gloria Romanorum
350-360
350-360
copy as LRBC II, 25 534
copy as LRBC II, 25
66
350-360
364-378
2098
W6
347-348
347-348
347-348
347-348
347-348
347-348
347-348
347-348
347-348
347-348
347-8
347-348
347-348
347-348
347-348
346-350
LRBC I, 455
LRBC I, 140a
as LRBC I, 140
as LRBC I, 148
as LRBC I, 158
LRBC I, 267
as LRBC I, 145
as LRBC I, 158
346-350
447
51
567
257
473
153
196
73
348
52
178
54
545a
372
77
576
214
as Bastien 21
as Bastien 33
RIC 189
2004, F213
2004
copy as LRBC II, 25
copy as LRBC II, 25
copy as LRBC II, 25
copy as LRBC II, 25
copy as LRBC II, 25
copy as LRBC II, 25
copy as LRBC II, 25
copy as LRBC II, 25
copy as LRBC II, 25
copy as LRBC II, 25
363
574
50
14
432
205
246*
305*
236*
38
47
109
134
136
173
206
copy as LRBC II, 25 516
copy as LRBC II, 25 423*
as LRBC II, 78
350
FOSSE LANE 1996: SPECIALIST REPORTS
Reign
Valentinian I
Reverse
Gloria Romanorum
Date
364-378
Valentinian I
Valentinian I
Valentinian I
Valentinian I
Valentinian I
Valentinian I
Valens
Valens
Valens
Valens
Valens
Valens
Valens
Valens
Valens
Valens
Valens
Valens
Gratian
Gratian
Gratian
Gratian
Gratian
Gratian
House of Valentinian
House of Valentinian
House of Valentinian
House of Valentinian
House of Valentinian
House of Valentinian
House of Valentinian
House of Valentinian
House of Valentinian
House of Valentinian
House of Valentnian
House of Valentinian
House of Valentinian
House of Valentinian?
Arcadius
Arcadius
Arcadius
Arcadius
House of Theodosius
House of Theodosius
House of Theodosius
House of Theodosius
House of Theodosius
House of Theodosius
House of Theodosius
House of Theodosius
Gloria Romanorum
Gloria Romanorum
Gloria Romanorum
Gloria Romanorum
Securitas Reipublicae
Securitas Reipublicae
Gloria Romanorum
Gloria Romanorum
Gloria Romanorum
Gloria Romanorum
Securitas Reipublicae
Securitas Reipublicae
Securitas Reipublicae
Securitas Reipublicae
Securitas Reipublicae
Securitas Reipublicae
Securitas Reipublicae
illegible
Gloria Romanorum
Gloria Romanorum?
Gloria Novi Saeculi
Gloria Novi Saeculi
Gloria Novi Saeculi
Gloria Novi Saeculi
Gloria Romanorum
Gloria Romanorum
Gloria Romanorum
Gloria Romanorum
Securitas Reipublicae
Concordia Auggg
Gloria Romanorum
Securitas Reipublicae
Gloria Romanorum
Securitas Reipublicae
Securitas Reipublicae
Securitas Reipublicae
364-378
364-375
364-378
367-375
364-378
364-378
364-378
364-378
364-378
364-378
364-378
364-378
364-378
364-378
364-367
367-375
367-378
364-378
367-375
364-378
367-375
367-375
367-375
367-375
364-378
364-378
364-378
364-378
364-378
378-383
364-378
364-378
364-378
364-378
364-378
364-378
364-378
364-378?
388-402
388-402
388-402
388-402
388-402
388-402
388-402
388-402
388-402
388-402
388-402
388-402
Victoria Auggg
Victoria Auggg
illegible
illegible
Salus Reipublicae
Salus Reipublicae
Victoria Auggg
Victoria Auggg
Victoria Auggg
Victoria Auggg
Victoria Auggg
illegible
Context
2062
2095
Ref
as LRBC II, 94
S.F. No.
49
as LRBC II, 279
as LRBC II, 279
LRBC II, 296
LRBC II, 321
as LRBC II, 96
as LRBC II, 96
as LRBC II, 92
as LRBC II, 282
as LRBC II, 478
64
326*
75
429
145
212
565
417*
222
249
56
152
349
359
127
207
132
545
133
369
223
224
300*
450
113
137
213
237*
147
307*
592
323*
168
361
371*
453
218
188
158
303*
440
510
154
244*
251*
62
63
393
533
575
as LRBC II, 97
as LRBC II, 97
as LRBC II, 97
as LRBC II, 97
LRBC II, 277
LRBC II, 303
LRBC II, 528
LRBC II, 339
2029
2001
2004
2029
2062
2024
2022
as LRBC II, 503
as LTBC II,503
as LRBC II, 503
as LRBC II, 517
as LRBC II, 78
as LRBC II, 78
as LRBC II, 92
as LRBC II, 92
as LRBC II, 96
as LRBC II, 269
as LRBC II, 275
as LRBC II, 276
as LRBC II, 279
as LRBC II, 280
as LRBC II, 280
as LRBC II, 481
as LRBC II, 164
LRBC II, 392
3013
2005, F223
2005, F223
as LRBC II, 796
as LRBC II, 1105
as LRBC II, 389
as LRBC II,389
Twenty one illegible 3rd to 4th-century coins were recorded: SF nos, 78, 217, 352, 353, 355, 362, 364, 367,
428, ?281, 443, 452, 456, 514 (found in layer 3010 structure 6 SW room), 531, 532, 542, 543, 544, 569 and
593
Thirty three illegible 4th-century coins were recorded: SF nos 40, 58, 85, 90, 139, 177, 181, 191, 193, 232,
279, 317* (from layer 2031),426, 427, 441, 444, 445, 446, 449, 470, 517 (layer 3013, Structure 6, SE
room), 535, 536, 537, 538, 539, 540, 566, 570, 571, 572, 579 and 581
W7
SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011
Six illegible 3rd to 4th-century copy coins were recorded: SF nos 82, 97, 209, 269, 414, 418
Three illegible 4th-century copy coins were recorded: SF nos 34, 57 and 135
Twenty one 4th-century copies were recorded: SF nos 116, 123, 156, 161, 184, 186, 190, 204, 208, 210,
220, 221, 230, 247, 316* (from layer 2031), 360, 373, 415* (from F104), 457, 471 and 583
Finally two post-Roman coins were recorded: A French jetton, late C14-early C15; Mitchiner 489; SF 578;
and a C19-C20 halfpenny, SF 547
ROMAN SMALL FINDS Lynne Bevan with a
contribution by Roger Tomlin
Introduction
The small finds collection is described below by
material rather than by function, given the relatively
small number of objects. Analysis of the assemblage
as a whole allows some trends and patterns to be
identified. Relatively few fixtures and fittings were
present, although a fragment of shale inlay from
furniture or a tray (Fig. 29.9), may indicate that luxury
goods were either present or being made on or in the
vicinity of the site. The wall plaster suggests the
existence of at least one well-decorated room in a
relatively sophisticated building. The small tool
assemblage – three chisels, a wedge, two ox goads
and a bucket handle – is much as one would expect
from an agricultural cum industrial community. The
three styli together perhaps suggest the record keeping
associated with the corralling and selling of cattle on
site, as has been mooted to have occurred at sites such
as Barnsley Park, Hambledon, and Rocester,
Staffordshire (Ferris and Cooper 1996, 149). The
relatively large number of items of jewellery –
particularly the ten pins and 16 bracelets – may be of
some particular significance as has been suggested
for the larger but similarly biased assemblages from
Uley, Lydney Park and Great Witcombe (Bevan
1998, 86, 88–9), while the inscribed silver ring
strikes a poignant chord even today. The lead
ossuary is a relatively rare find on a Romano-British
site, and its presence adds another example of an
unusual burial rite being practised in Roman Shepton.
Catalogue of copper alloy objects
Seven items, cat nos 1, 8, 9, 10, 25, 26, 27 are
illustrated (Fig. 29)
1
W8
Pin, broken. The globular head was covered with
close-set radiating diagonally incised grooves
and the neck was delineated by approximately
ten horizontal bands. This is similar to a pin head
from a post-Roman posthole at Lion Walk,
Colchester (Crummy 1983, fig. 31.499, 30–31).
SF 166, 110/255, east of Structure 12. Fig. 29.1.
2
Pin, complete, with a small spherical head, a
common type of pin which was used throughout
the Roman period (Cool 1990, fig. 1.5, 151–2).
SF 30, F131.
3–7 Pin fragments: 3, SF 357, 60/305, Structure 6;
4, SF 411, 130/340, Structure 5; 5, SF 115, 75/
295, Compound 7; 6 SF 91, 80/265, Compound
8; 7 SF 95, 80/280, Compound 8.
8
Bracelet fragment, D-shaped section, broken at
both ends, with intricate decoration in the form
of a thick raised area of deep grooves and
horizontally hatched vertical bands at the front
of the armlet which narrows at the shoulder
where it is decorated with a vertically hatched
horizontal band. SF 381, 125/335, Structure 9.
Fig. 29.2.
9
Bracelet fragments, D-shaped section with
decoration consisting of diagonal grooves which
give way to vertical grooves at one broken end.
SF 378. Fig. 29.3.
10 Bracelet of double-stranded twisted wire with a
D-shaped section broken at both ends. SF 39,
110/375, east of Structure 2. Fig. 29.4.
11 Bracelet with three ring and dot motifs at one
end and an incised line along each edge. SF 290,
layer 2026, Structure 8.
12 Bracelet fragment, rectangular-sectioned with
decoration in the form of a continuous wave
motif enclosed by bands, broken at both ends.
SF 159, 100/290, Compound 8.
13 Bracelet fragment, flat with a cut-out crenellated
edge, and a perforated terminal at the unbroken
end. SF 194, 125/260, east of ditch F236.
14 Bracelet fragment, flat with a cut-out crenellated
edge. SF 163, 110/255, Compound 8.
15 ?Wire bracelet. Hooked fragment with squareshaped section and traces of incised diagonal
striped decoration on upper face. SF 128, 75/
300, Structure 17.
16 ?Bracelet fragment. Strip with traces of diagonal
and linear decoration. SF 586, unprovenanced.
17 ?Bracelet fragment, curved strip. SF 332, pit
F253.
FOSSE LANE 1996: SPECIALIST REPORTS
Fig. 29 Copper alloy (1-7) and shale (8-10) objects
18
19
20
Bracelet fragment, traces of incised designs on
outside face. SF 93, 60/405, west of Structure 1.
Finger ring, broken curved fragment widening
at shoulder. SF 412, 130/340, Structure 9.
Ear-ring pendant. Small segment of wire formed
into a rod with chainlink loops at either end,
probably a pendant from an ear-ring which once
held a glass bead. A pair of similarly sized
chainlink ear-ring pendants with dark blue beads
in situ was found in a 4th-century context in
Colchester (Crummy 1983, fig. 53.1797). SF 3,
70/390, road F101.
W9
SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
W 10
?Wire bracelet. Two strands of twisted wire.
SF 425, layer 2098, Structure 9.
Hooked object, broken, with oval area decorated
with circular motif. SF 155, 100/255, east of
Structure 12.
Bracelet or necklace hook. SF 523, Compound
7.
Tweezers with flared blades, one broken. SF 141,
100/270, Structure 11.
Horse harness. Long, hexagonal stud with a
raised oval boss with deep median groove. Two
disc-headed shanks protrude from the reverse.
This type of stud, which was probably used as
horse harness, has been found on many Roman
sites (Allason-Jones and Miket 1984, 3.870,
3.871, 237). SF 104, 55/310, Structure 6. Fig.
29.5.
Plate, diamond shaped with raised, split panels
at each end and a flat central area with four
decorative pierced holes. The reverse has two
studs for the attachment of the plate to a strap.
SF 110, Structure 6. Fig. 29.6.
Mount, shield-shaped, leaded copper alloy,
decorated with three deep grooves. SF 529;
unprovenanced. Fig. 29.7.
Stud, circular rosette shape, crudely decorated
with a series of irregular incised lines around
the outer circumference, and traces of iron from
a central attachment, now corroded. A number
of similar studs recovered from 1st-century
contexts at Colchester were used as military belt
and apron fittings (Crummy 1983, fig. 151:4204,
4205). SF 211, 135/295, east of F236.
Small ring, leaded copper alloy, D-shaped
section, probably from clothing or horse harness.
SF 118, Structure 6.
Ring fragment, square section. unprovenanced.
Ring fragment, D-shaped section. SF 4, 65/385,
south of road F101.
Stud, oval, possibly from a small box or item of
furniture. SF 16, 60/400, west of Structure 1.
Strip, rectangular with uneven ends and three
incised lines, rectangular-sectioned. SF 526,
unprovenanced.
Rod, broken, square-sectioned. SF 336, pit F253.
?Vessel fragment. Plate with grooved slightly
raised edge. SF 239, ditch F223.
?Box or furniture fitting. Circular pierced
terminal. SF 313, ditch F245.
Terminal, rounded end decorated with four lines
of incised dots, leaded copper alloy. SF 587,
unprovenanced.
Fitting. Tapering curved object, leaded copper
alloy. SF 521, beneath Structure 7.
Ferrule divided into three bead-shaped segments
by two grooves. SF 588, unprovenanced.
Plate fragment, rectangular, broken, one straight
end, serrated edges. SF 298, pit F225.
41–4 Wire fragments. 41 SF 585, unprovenanced;
42 SF 144, 105/280, east of Structure 11; 43
SF 335, pit F253; 44 SF 524, pit F335.
45–51 Rod and strip fragments. 45 SF 201, 115/280,
east of Structure 12; 46 SF 262, 140/290, east
of F236; 47 and 48 SF 522, F335; 49 SF 26,
F133; 50 SF 291, ditch F223; 51 SF 324,
unprovenanced.
52–8 Sheet fragments. 52 SF 29, 85/375, Structure 2;
53 SF 125, 95/260, Structure 12; 54 SF 53, 120/
365, Structure 4; 55 SF 274, 145/290, east of
F236; 56 SF 344, layer 1028; 57 SF 305, bank
F234; 58 70/390, road F101.
Silver ring Roger Tomlin
Fragment of a silver ring (not illustrated) , broken at
the shoulder with an intact bezel, originally c 22mm
in diameter. An inscription which reads ‘ME l MOR’,
probably memor, ‘mindful’, is enclosed within the
round bezel which measures 8mm in diameter. There
is an incised roundel below the legend, which
corresponds to the two roundels on each shoulder of
the bezel. Memor is a cognomen, and even a cult-title
of Minerva at a shrine in North Italy (cf ILS 2603, a
votive brought back from Britain), but here it is
probably an adjective understood by the donor of the
ring, presumably, and by its wearer: either (sis) memor
(mei), ‘Remember me’, or (sum) memor (tui), ‘I
remember you’. The latter is perhaps to be preferred,
in view of the rings inscribed MEMINI TVI (ClL xiii
10024. 71(b) and (c), ‘I remember you’), and MEMINI
TVI MEMINI ET AMO (ibid, 72, cf 73, ‘I remember
you; I remember and I love (you)’). SF 296; layer
2029, floor Structure 7.
Iron objects (none illustrated)
1
2
3
4
5
6
7
8
Part of a buckle. SF 111, Area B.
Pin shank. SF 345, grave F208.
Fragment from a stylus with a broad-based
eraser, broken across shank and very corroded.
SF 341, grave F150.
Fragment from a stylus with a broad-based
eraser, broken across shank and very corroded.
SF 108, 80/270, north of Structure 8.
Stylus with a rounded eraser, very corroded.
Unprovenenaced.
Chisel, similar to a possible Iron Age example
from Hod Hill, Dorset (Manning 1985, plate
11:B43, 24). 75/265, Structure 8.
Chisel or punch, similar in size and shape to a
mid 1st-century example from Hod Hill, Dorset
(Manning 1985, Plate 5:A23, 10). SF 151, 110/
270, east of Structure 11.
Chisel or punch, similar in size and shape to
another mid 1st-century example from Hod Hill,
FOSSE LANE 1996: SPECIALIST REPORTS
9
10
11
Dorset (Manning 1985, Plate 5:A25, 10). SF
143, 105/255, east of Structure 12.
?Wedge. 120/365, Structure 4.
Bucket handle terminal consisting of a looped
circular-sectioned length of iron rod, similar in
style to an Iron Age or mid-first century example
from Hod Hill in Dorset (Manning 1985, Plate
47:P14, 103). SF 419, 80/375, west of Structure
2.
Conical object attached to a circular disc of
copper alloy. SF 276, ditch F230.
A total of six shoe cleats were found: SF 315, 2025,
Structure 7; F133; south of F101; Ditch F134;
Structure 2 and unprovenanced. Thirty hobnails were
recovered. A total of 15 came from grave F150 with
the remainder as surface finds, ten of which came from
the areas east of Structure 12, and one each from
Structures 1, 2, and 4, road F101 and east of F236 in
Area B. Two ox goads were found as surface finds in
the area of Structures 4 and 10. A number of
miscellaneous items were also found comprising a
length of iron rod with a looped end, a hook, two
staples, three rings, seven fragments of rod and wire,
21 fragments of plate and over 300 nails.
Lead objects (none illustrated)
The total weight of the lead objects was just under
4kg.
1
Lead ossuary, circular drum-shaped container,
originally equipped with a domed top and central
spout which survives as a complete base and a
number of small fragments from the sides, top
and spout which were severely degraded beyond
reconstruction. This small, undecorated ossuary
with its integral spout is an example of a ‘pipe
burial’, a Roman burial custom known from Italy
and the Rhineland and elsewhere (for discussion
see Philpott 1991, 28), involving the pouring of
libations down a channel protruding above
ground level into a coffin or ossuary. The
channel was made from various materials, which
had usually been re-used or adapted, such as two
vertical imbrices leading into a tile cist at St
Pancras, Chichester (Down and Rule 1971,
grave 323), and the re-used amphora spouts from
Osola Sacra, Ostia which would also have been
visible as grave markers (Calza and Becatti
1977, 69). At Falerone in Italy, a vertical lead
pipe was attached to a small stone sarcophagus
containing a cremation-burial and various grave
goods (Wheeler 1929, 4).
In Britain, a pipe burial involving a lead pipe
is known from Creffield Road, Colchester
(Wheeler 1929, 4), and a lead pipe which
probably fulfilled a similar function was found
with a cremation at Mancetter, Warwickshire
(Booth 1982, 134–6). The closest parallel is
from Caerleon (Wheeler 1929), where a slightly
larger drum-shaped ‘canister’ decorated with
three reel-patterned bands and with a lead pipe
attached to the centre of its lid was found in a
stone-lined cist. The pipe had originally
protruded beyond the cist and above the former
land surface. The burial of the canister, which
contained the cremated remains of a mature
male, was Hadrianic or later.
Philpott has suggested that pipe burials resulted
from ‘a strong desire... (that) the dead should
have direct contact with the living and that
perhaps such libations were seen to nourish the
dead in a particularly vivid way’ (Philpott 1991,
28). The act of libation might also have been
viewed as a placatory or appeasing gesture
towards the spirits of the deceased, perhaps
combined with a desire for guidance or the
bestowal of good fortune from the other world.
The distribution of ossuaries was concentrated
in the main military and urban centres, and the
occurrence of ossuaries on relatively rural sites
such as this example is rare and ‘may indicate an
adoption of the ossuarium by high-status romanised
natives or the burial of immigrants on their estates’
(Philpott 1991, 28). The large quantities of jewellery
recovered from the site would certainly support
the presence of high-status individuals, the majority
of whom were probably members of the local,
Romanised population. Seen in this context, the
adoption of a specifically Roman burial practice is
less surprising.
Base diam 190mm, th 3mm, wt of fragments
1188gm. SF 380, F242 (Fig. 10).
2
Caulking. T-shaped fragment of caulking made
from a thick, rounded core of lead fused with
folded sheet. 60/285, Structure 14.
3
?Weight. Roughly conical object. SF 394,
unprovenanced.
4
Pipe collar. Collar with furled edge, diam
105mm. SF 33, unprovenanced.
5–6 Circular objects, one with central perforation,
both unprovenanced.
7
Rounded fragment with flattened base.
unprovenanced.
A total of 10 pot cramps were found, five still
attached to fragments of pottery, two Black Burnished
ware vessels, two sandy reduced coarseware vessels,
a Samian vessel and a greyware vessel. The located
items came from east of F236 (5), Structure 9, east
of Structure 7, east of Structure 11, and east of
Structure 12.
A total of 13 lead offcuts were found as surface
finds, with provenanced items coming from
Compound 6 (2) Structures 4, 6, 11 and 17, road F101,
and east of F236. Numerous small waste globules
from lead working were found.
W 11
SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011
Worked bone (not illustrated)
1
2
Pointed pin fragment. SF 253, F224.
Pointed pin fragment. SF 306, F243.
Shale
There is no evidence from this or the 1990 excavation
(Leach 2001) for the on site manufacture of shale
items. It is, however, possible that no 1 (Fig. 29.9)
may have been manufactured or modified on site. The
source of the shale is likely to have been Kimmeridge
from which shale was exploited at a number of
contemporary sites in Purbeck (Sunter and Woodward
1987). The standard of preservation of the items was
generally good apart from some delamination at the
edges of no 1, which might have been damaged and
subsequently discarded during the manufacturing
process.
1
2
3
W 12
Flat fragment of shale. Two motifs are visible
on its upper surface; a single spiral and an ‘S’shape, both enclosing incised dots. A fragment
has been cut from the upper surface The
spiralform decoration is reminiscent of motifs
on later Iron Age pottery, in contrast to the more
formalised and repetitive decoration seen upon
published examples of trays from other Roman
sites such as Colchester (Crummy 1983, fig.
75:2022, 2023, 69 and 71), Silchester (Lawson
1975, figs 11.87–9, 12.90), and similar
fragments described as ‘plaques’ and ‘tablets’
from Norden near Corfe Castle, Dorset (Thomas
1987, figs 19 and 20.18–22, 30–35). However,
similar motifs combined with more regular
panels of intricate geometric banding have been
identified on an almost complete tray or
‘trencher’ found in a cremation burial from
Grange Road, Winchester which had vessels, a
spoon and knives, and two pork joints on it
(Biddle 1967, fig. 6.233). Although Thomas has
suggested that such items might also have been
use for other purposes including ‘wall plaques’
and ‘small table tops’ (Thomas 1987, 33), that
the Winchester object was used as a tray appears
beyond doubt. The dating evidence suggests a
general 1st to 2nd-century date for these items
(Biddle 1967, 248–59; Thomas 1987, 35), which
might reflect a fashion in tableware and serving
etiquette. The cut-out section on the decorated
face of this fragment might indicate that shaleworking was carried out on the site. SF 6, 85/
405, Structure 1. Fig. 29.9.
Spindlewhorl, elliptical section. Fig. 29.8.
Two joining oval-section, lathe-turned fragments
from a large bracelet. The item has three grooves
on the exterior and a slightly raised bevel on the
4
inside, which might have been designed to hold
the bracelet to the upper arm, perhaps over
clothing. The armlet is an exceptionally large
piece with an internal diameter of 120mm;
similarly sized armlets have been recorded at
Norden (Thomas 1987, 30) where internal raised
bevels were also recorded. SF 27 and SF 31;
F133. Fig. 29.10.
(not illustrated) Three joining fragments from a
child’s bracelet with a D-shaped section. F160.
Wall plaster
Four fragments of wall plaster were recovered, all of
which had retained traces of pigment. A similarly
small group was noted in 1990 (Morgan 2001, 230).
The pigments ranged in colour from dark red to light
pink. As no traces of painted decoration were visible,
the fragments appear to have originated from plain
painted borders rather than from decorated panels.
None are illustrated.
1
2
3
4
Fragment with red/brown paint. SF 518, F224.
Fragment with bright pink paint. SF 563, F225.
Fragment with traces of pale pink paint. SF 24,
60/310, Structures 6 and 17.
Fragment with traces of pink paint. SF 84, 80/
260, Structure 8.
THE BROOCHES Donald Mackreth
Where a brooch in the present assemblage is also
represented in the 1990 report, the reader is referred
to the discussion there, the conclusion only being
repeated here (Mackreth 2001). All the brooches,
except nos 22 and 23 (of iron), are made from copper
alloy; none are illustrated.
Colchester Derivatives
Brooches 1–3 have or had their springs mounted in
the Polden Hill manner: the axis bar through the spring
is mounted in a pierced plate at the end of each wing,
the chord being held either by a rearward-facing hook
or passing through a pierced crest on the head of the
bow.
1
Each wing has a relieved beaded ridge separated
by a wide flute. The bow has a beaded pseudo
hook dying away in the V formed by two more
beaded ridges. On each side of the head of the
bow is a plain additional moulding. The foot has
a slight projection and the catch-plate a small
triangular piercing. SF 41, surface find area B.
Of the same family as Brooches 7 and 8 in the
first report, the conclusion there was that the
date-range of the Polden Hill version of the
FOSSE LANE 1996: SPECIALIST REPORTS
2
3
general family ran from c AD 65 to about 125,
and that those dating to perhaps as late as AD
125 may well have been survivors in use.
The head is missing. In the panel on the upper
bow, which is outlined by a groove, is the trace
of at least one cell, and perhaps of three. Beneath
the panel is a slight waist, then two lenticular
bosses above a median arris. The bow tapers to
a narrow projecting foot. The catch-plate is solid.
SF 281, layer 2017.
The family to which this brooch belongs has
both the Polden Hill spring system, as this
example would have had, as well as hinged pins.
Discussed under Brooches 12–15 in the first
report, the date-range for those with Polden Hill
spring systems is mainly later 1st to mid 2nd
century.
The wings are plain. The bow has a very short
additional moulding on each side, and a relieved
ridge down the upper part stopping at a marked
cross-flute with a raised border top and bottom.
The rest of the bow is plain and ends in a small
projecting foot. The catch-plate is solid. SF 258,
F223.
Although not a member of a specific group,
the additional mouldings and the ornament in
the middle of the bow show affinities with the
brooches which have passed in review and a
date-range of c AD 75–125/50 is likely.
top but rapidly tapers to a thin foot marked by
two projections. On the top of the bow is a single
short ridge dying away into the V formed by a
pair of others meeting in a point below. The
catch-plate is solid. SF 330, pit F254
Very like Brooches 25–6 in the first report, this
example probably dates from the later 1st
century to AD 150/75.
Unclassified
7
Trumpet
8
Brooches 4–7 have hinged pins.
4
5
6
In two parts and badly corroded, the surviving
wing is plain and only the upper bow survives.
This has a flat back, a moulding apparently on
each side of the head and a broad one down the
middle. SF 130, 105/255, east of Structure 12
There is little to go on and a general date
running from the later 1st century to about AD
175 may be suggested.
Each wing has a sunken moulding at its end. On
the head of the bow is a tab with a circular
depression. The bow has a step down each side
of the upper part and a projecting foot. Between
the lower end of the stepped part are two
lozenges each with a central cell for enamel, now
missing, and a bordering groove. On each side
between the lozenges is a small ‘gablet’. The
catch-plate is solid. SF 301, layer 2022
Of the same family as nos 44–8 in the first
report (Mackreth 2001, 191), more specifically
nos 46–7, the dating arrived at there has not
improved since the writing of that report: later
1st to AD 200 by which time any in use were
survivors as manufacturing had ceased as near
as can be assessed for all Colchester Derivatives
by c. AD 175.
Each wing has two reels. The bow has a wide
There is a ridge running across the head of the
bow and the plain wings. The bow is very broad
at the top tapering to a wide foot which projects
boldly under a step. The profile is distorted, but
once had an almost flat area above the main face
which has a deep groove on each side and a small
central ridge at the very top. SF 81, 55/305,
Structure 6
In very broad terms, the overall proportions are
like those of Brooches 34–8 of the first report
(Mackreth 2001, 189–90), but none there is
obviously of the same decorative type. However,
the conclusion reached is almost certainly
applicable here: there are no grounds for a 1stcentury date, nor any evidence that any should
be expected in the 3rd.
The pin is hinged, the axis bar being housed in
a narrow projection behind the bottom of the
trumpet head. There are the remains of a caston loop on a tall pedestal having two grooves.
The trumpet head is very broad and shallow at
the top, with a deep sweep on each side to a
width less than that of the lower bow. The knop
is of the usual petalled form with a single ridge
top and bottom and all moulded on the front of
the bow. The lower bow is broad, has a median
arris and tapers slightly to the triple moulded
foot. SF 140, 100/275, Structure 11
Although few appear to come from the same
moulds, there is a family likeness which, when
coupled with the distribution, which lies in the
counties from Hampshire to Somerset, points to
a distinct variety. However, none is dated and
only a general range lying in the 2nd century
can be suggested, there is little in the design to
suggest that it could have started in the 1st
century.
Strip
9
The head of the bow is rolled under to house the
axis bar of the hinged pin. There is effectively
no head-plate, but one is defined roughly by the
projecting arcs housing the eyes which are made
up of a dot and circle. The lower bow with the
W13
SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011
catch-plate is missing. The bow as it survives
has a broad flute down each side and a groove
down the middle across which lies a line of
horizontal lines. SF 409, F262
An almost exact parallel occurs in the first report,
Brooch 82, except that has a rolled over head and
was better related to the Alesia – Hod Hill sequence
(Mackreth 2001, 197). Here the associations, when
married to the patently non-continental way of
conveying the impression of there being a bead-row,
are entirely with the Strip and, as such, is an example
of that type deriving from members of the pre-Aucissa
sequence. The determining feature is the presence of
eyes. No Aucissa proper has these and it is clear that
the general family had abandoned them well before
the conquest. This brooch may have survived in use
as late as that, but hardly beyond.
Fragments
10
11
12
13
14
15
16
17
18
19
20
21
22
23
The lower bow and catch-plate of a brooch of
unknown type. Probably before AD 250. SF 530,
unprovenanced
Complete pin and bilateral spring of five coils
probably once mounted between a pair of
pierced lugs. Second century. SF 175, 125/255,
east of F236
Hinged pin. SF 20, unprovenanced
Hinged pin. SF 525, unprovenanced
The end of a pin possibly from a brooch. SF
268, 60/280, area of Structure 14
Hinged pin. SF 238, 130/275, east of F236
Hinged pin. SF 245, F224
The lower bow and catch-plate of a brooch of
unknown type. Probably before AD 250. SF 1,
60/395, road F101
Spring fragment. SF 242, 130/275, east of F263
Penannular brooch with out-turned terminal end.
SF 22, 95/375, east of Structure 2
Spring fragment. SF 332, F253
Hinged pin. SF 416, 130/340, Structure 9
Iron strip brooch with hinged pin. SF 70, 35/
315, Compound 6
Hinged pin and strip from iron brooch. SF 375,
F258
THE GLASS Birgitta Hoffmann
A total of 117 Roman glass fragments was found.
The fragments can be divided by type as follows:
cast: 1; monochrome: 4; colourless: 11l greenish:
28; blue-green: 33; bottle: 19; window glass: 12;
and objects: 9.
The material derives mainly from the 1st and 2nd
centuries and the late Roman period, with diagnostic
finds from the later 2nd and early 3rd centuries
W 14
(cylindrical beakers etc) being absent. None of the
material is of particularly high quality.
Pillar-moulded bowls, like no 1 (Fig. 30.1), are an
extremely common find on 1st-century settlement
sites. They can be found, for example, amongst the
finds from the military works depot at Longthorpe
(Dannell and Wild 1987, 51) and at Kingsholm (Cool
and Price 1985, 45 nos 4–9), both of which were
occupied in the (Claudio-) Neronian period, as well
as in most southern British towns eg Silchester (Price
1984, 117, nos 1 and 2). At Vindolanda, however,
which starts in c AD 85, none were found amongst
the nearly 4000 glass fragments from the vicus
(Hoffmann forthcoming). Three further examples of
deep-blue pillar-moulded bowls are known from the
1990 Shepton Mallett excavations (Price and Cottam
2001, nos 1–3a).
Two parallels are known for no 3 (Fig. 30.3), a dark
brown fragment from Dorchester, Greyhound Yard
(Cool and Price 1993, fig. 84.23) and a blue/green
fragment from Usk (Price 1995, fig. 44.68). In both
cases the original shape of the vessel can not be
reconstructed, but the strong colour might point to a
1st-century date. The high percentage of brown or
strong yellow/brown glass in the assemblages from
both the 1996/7 (3 fragments) and 1990 sites (5
fragments) is remarkable, and appears to outweigh
any other strong colour from the site.
Number 4 (Fig. 30.4) comes from a globular flask
with wheel-cut lines. This type of bottle is particularly
common in the later 3rd and 4th centuries. However,
the narrow neck with its very late starting funnel seem
to make the vessel more likely to belong to an earlier
tradition (Cool and Price 1995, 149, fig. 9.3).
Decorated flasks of this type are unusual in Britain,
but examples are known from Libya (Price 1985, 78;
99, no 50, fig. 6.4), Kisselbach/Hunsrück (dated to
the end of the 1st century (Hopstätter 1942)), KölnSeverinskloster (dated to c. AD 200) and
Luxemburger Straße (Fremersdorf 1984, 42, no 107),
and from Ladenburg (Hoffmann 1996, 162, kat no
L99–100). Number 5 (Fig. 30.5) might be part of one
of the 2nd-century drinking vessels with linear cut
lines, although not enough survives to be certain.
These vessels are known in a number of different
shapes and can have either pad bases or pushed-in
base-rings, with other base types occurring more
occasionally (Cool and Price 1995, 79–82).
Handles like nos 6 (Fig. 30.6), 7 and 8 and the brown
rim no 2 (Fig. 30.2), are most likely to have come
from long-necked conical (or less likely globular)
jugs. These occur in a variety of colours, with blue/
green being by far the most common (Cool and Price
1995, 120–3) and range in date from the late Flavian
period (Cool and Price 1995, 122) to about AD 150–
170 (Price 1980, 66, fig. 15.6 and 7; Price 1987, 204,
fig. 3.20; Price and Cottam 1994, 225, fig. 104.9 and
10). The handles found represent probably three
FOSSE LANE 1996: SPECIALIST REPORTS
Fig. 30 Vessel glass (1-9), bangle (10) and beads (11-13)
different vessels and another four were recovered
from the 1990 site. Such a high number of vessels of
this type should come as no surprise as they are a
common find on sites of the later 1st and 2nd
centuries, with examples known from Colchester,
Verulamium, Gloucester (Cool and Price 1995, 123),
Carlisle, 19 examples (Price 1990, 174; Cool and Price
1991, 166; Price and Cottam 2001, 240) and
Vindolanda, at least 17 examples (Hoffmann
forthcoming).
W 15
SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011
Indented vessels, cups, bowls and flasks were in
use from the 1st to the 4th century, although too little
survives of no 9 to allow any statement as to the
original shape of the vessel. The same holds true for
no 11 (Fig. 30.7), whose inturned rim is most likely
to have come from a globular or indented jar. These
vessels were fairly common in the 1st and 2nd
centuries in Britain (Cool and Price 1995, 112),
although the diameter of this example, at c 100 mm,
makes it one of the larger examples of the type.
Number 10 belongs to a globular bath flask with
‘dolphin handles’, which is a frequent shape on sites
from the 1st to the 3rd century (Cool and Price 1995,
156).
Bottles, probably four-sided (although other shapes
do occur), are represented by nos 15–18. Bottles are
usually easily identifiable by their characteristic thickwalled fragments. They form a major percentage of
any late 1st and 2nd-century assemblage and the 19
fragments found in 1996/7 are quite in keeping with
the 30 found on the 1990 site. These vessels were
mainly used for storage purposes and occur in many
different sizes (ranging in capacity from c 0.2 to
several litres). Number 18 probably comes from one
of the smaller bottle types. Number 17 is a small
fragment of one of the base marks that are typical of
these vessels, in this case circles, for which a wide
range of combinations are known. Such bottles date
mainly from the mid 1st to the 2nd or early 3rd
centuries although they can occasionally be found in
4th-century assemblages (Cool and Price 1995, 184).
Fourth-century glass is represented by nos 12 and
13. These curved rims are typical of a number of late
Roman conical and hemispherical vessels, although
too little survives here to identify the original types.
Similar rims can be found in most late Romano-British
assemblages as at Silchester (Price 1984, 118, no 3).
Thirteen conical beakers with this rim type were found
on the 1990 site, and a further 30 specimens have
been listed from elsewhere in the region, for example
from Bath, Cirencester, Ilchester, Dorchester and the
rural sites of Frocester Court and Catsgore (Price and
Cottam 2001, nos 28–31; 33; 35–6; 29–41 and 42–45c).
The other certainly late Roman type is the greenish
vessel with fire-rounded rim, no. 14 (Fig. 30.8) and
three other similar such rim fragments were found in
1990 (Price and Cottam 2001, nos 38–40). Similar
vessels have been dated elsewhere to the second half
of the 4th to the early 5th century, with the most
famous examples coming from the glass hoard at
Burgh Castle (Harden 1983, 82–3, nos 85–9, fig. 37).
The large bangle fragment no 19 (Fig. 30.10)
belongs to Kilbride-Jones’ Class I (Kilbride-Jones
1937–1938, 367–72; Stevenson 1956, 208, 218;
Stevenson 1976). The applied foil over a usually
bluegreen body is characteristic, as is the piece’s rather
‘heavy’ character. Close parallels have been found at
Traprain Law, the Roman fort of Camelon (Kilbride-
W 16
Jones 1937–8, fig. 1), the native promontory fort of
Mains of Ethie, Angus (Wilson 1980, 121) and other
sites in Southern Scotland. The southernmost piece
known previously comes from the Roman fort at
Vindolanda, however, where it was found in a
probably residual early 3rd-century context
(Hoffmann forthcoming), so its presence in Somerset
is somewhat surprising. In view of the finds from
Traprain Law, Kilbride-Jones dates this type to the
late 1st century (Kilbride-Jones 1937–1938, 367–72).
Beads of the same type as no 20 (Fig. 30.11) were
dated by Guido (1978, 96) to the 3rd and 4th centuries,
but more recent finds from Vindolanda (Hoffmann
forthcoming) and Strageath (Price 1989, 197, 202,
fig. 102,7) appear to belong to the 2nd century.
Although never found in large quantities, examples
can be found all over Britain and similar finds from
the same region are known from Cirencester, Bradley
Hill, Lufton and Mendip (Guido 1978).
Bluegreen segmental beads nos 21–23 (Fig. 30.12,
13) appear to be particularly common on later Roman
sites, eg Portchester and Lympne (Guido 1978, 201),
with a 3rd-century example now known from
Vindolanda (Hoffmann forthcoming). Other examples
come from Colliton Park/Dorchester, Cadbury Castle,
Hengistbury Head, Camerton and Ham Hill (Guido
1978, 200).
Bluegreen annular beads like no 24 tend to be found
on Roman rather than Iron Age sites but there is
currently no more precise dating for the type. Other
examples are listed by Guido under her group 6iib
(Guido 1978, 66, 143–5).
Number 25 is part of Guido’s Group 7v (Guido
1978, 172). This type is quite common in the region,
with examples known from Dorchester, Somerleigh,
Sydling, Cirencester, Silchester, Amesbury, Meare,
Glastonbury and Ham Hill (Guido 1978, 152–5).
There is also an earlier find from Shepton Mallet.
Numbers 26–27 are plain glass gaming counters,
which are a common find on Roman sites of all
periods. They are usually associated with board
games, although other uses, such as accounting and
calculation have occasionally been cited (Price 1995,
129). Black appears to be the most common colour
on most sites, usually far outweighing all other
colours, including white.
The window glass from the site reflects both broad
types known from Britain. The bluegreen cast
fragments (nos 28 and 29) represent the matt/glossy
variety that can be found on Roman sites from the
later 1st century, whilst the greenish, thin variety (nos
30–42), which is probably made by blowing a large
cylinder and then cutting it open, only came into use
in the later Roman period (Harden 1961). The latter
material is likely to be associated with the stone
buildings of Period 5.
In summary, the glass recovered gave a reasonable
overview of the glass used in everyday contexts in a
FOSSE LANE 1996: SPECIALIST REPORTS
settlement of the later 1st and first half of the 2nd
century, with, from the evidence of the glass, reoccupation in the 4th century. The present assemblage
fits well with the material recovered in 1990 and the
only unusual fragment is the Type I glass bangle
which, in view of its otherwise mainly southern
Scottish distribution, has to be counted as an import.
Catalogue of Roman glass
9
10
Cast glass
1
Rim fragment, bluegreen; one rib; no bubbles;
no weathering. RD: too small to measure: dims:
21 x 40 Th: 4 EVE: 0.4. SF 174, 110/280, east
of Structure 11. Fig. 30.1.
11
Blown glass
2
3
4
5
6
7
8
Brown
Rim fragment, brown; rim folded out, up and
in; no bubbles; no weathering. D: 30 mm PH: 6
Th: 3 EVE: 0.14. SF 636, F253. Fig. 30.2.
Body fragment, dark brown; folded out tube out
of body, definitely not base ring; no bubbles; no
weathering. D: 120 H: 6 mm Th: 2. SF 505, layer
1019. Fig. 30.3.
Also: 3a Dark brown body fragment, from
globular vessel; small bubbles; no weathering.
Dims: 25 x 47 Th: 2. SF 506, layer 1019.
Greenish
Nineteen fragments (some joining), greenish;
neck, body and base of globular flask with two
horizontal linear-cut lines, and two further ones
on the lower body; flat base; bubbles; no
weathering. PH(neck): 64 W(neck): 16
PH(body): 33 D(base): c.45 EVE: 0.6. SF 11,
75/405, Structure 1. Fig. 30.4.
Body fragment with linear cut lines; greenish;
?cylindrical vessel; no bubbles; no weathering.
Dims: 21 x 30 Th: 2 EVE: 0.2? SF 255, F224.
Fig. 30.5.
Bluegreen
Handle fragment with central rib; bluegreen;
elongated bubbles; no weathering. W(handle):
22–24 PH: 42 Th: 5 EVE: 0.14. SF 629, 105/
385, east of Structure 2. Fig. 30.6.
Lower handle attachment, bluegreen; handle
with central rib, drawn out over body and
pinched; elongated bubbles; no weathering.
W(ext): 43 PH: 36 TH: 7 EVE: 0.14. SF 72, 50/
275, east of Structure 7.
Handle fragment and lower attachment,
bluegreen, handle with central rib, drawn out
and pinched over the body; elongated bubbles,
greenish striations; no weathering. W(handle):
27–34 PH: 67 TH: 5 EVE: 0.17. SF 476, 125/
255 east of F236.
Perhaps from similar vessel: 8a One body
fragment, bluegreen. one straight rib; no
bubbles; no weathering. Dims: 18 x 24 Th: 1.
SF 509, 75/285, north of Structure 8.
Body fragment, bluegreen with indent;
bluegreen glass with little bubbles; no
weathering. Dims: 11 x 18 Th: 1 EVE: 0.14. SF
630, 110/255, east of Structure 12.
Handle fragment and shoulder of globular bath
flask; handle trails along shoulder, neck and
loops from the rim back onto shoulder and back
onto the rim; bluegreen; bubbles; no weathering.
H(ext): 34 Th: 5 EVE: 0.17. SF 98, 95/255,
Structure 12.
Two joining rim fragments, bluegreen, rim rolled
inwards, many bubbles and greenish striations,
no weathering. D: 100 H(ext): 24 Th: 1. SF103,
85/290, Compound 7, and SF 489, 75/290,
Compound 7. Fig.30.7.
Late Roman glass
12
13
14
Rim fragment, bluegreen/colourless; curved rim,
cracked off and left uneven; many small bubbles;
no weathering. D: 80 H(ext): 18 Th: 1 EVE: 0.2–
0.4. SF 282, 125/270, east of F236.
Rim fragment. Greenish; rim slightly turned out,
and cracked off straight; small bubbles; no
weathering. D: 80 PH: 12 Th: 2 EVE: 0.2. SF
508, 110/290, east of Structure 11.
Rim fragment, greenish; rim fire-rounded;
slightly outturned; body ?globular; many
bubbles, no weathering. D: 85 H(ext): 21 Th: 1
EVE: 0.4. SF 68, 120/345, Structure 9. Fig. 30.8.
Bottles
15
16
17
18
Rim fragment, colourless; small bubbles; matt
surface; rim folded out, up and in and flattened on
top. D: 45 H(ext): 7 EVE: 0.17? SF 636, F253.
Rim fragment, bluegreen; rim folded out, up and
in; many bubbles; no weathering. D: 52 H(ext):
21 Th: 5 EVE: 0.28. SF 5, 115/365, Structure 4.
Fig. 30.9.
Base of straight-sided bottle, bluegreen; small
bubbles; dull surface; base design: one circle;
not cast. Dims: 13 x 21 Th: 5 EVE: 0.14. SF
599, 105/360, west of Structure 4.
Reeded handle fragment, bluegreen; many
elongated bubbles, greenish striations; no
weathering. W(handle): 35 mm H(to turn): 33
Th: 2 EVE: 0.14. SF 596, pit F256.
Glass objects
19
Bangle
Bangle fragment (60 degrees); greenish glass
W 17
SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011
20
21
22
23
24
25
26
27
with three uninterrupted and three interrupted
opaque yellow trails on the outside. D(inner):
60 H: 18 Th: 13. SF 331, pit F253. Fig. 30.10.
41
42
Beads
Fragment, dark blue; rectangular bead with
facets at the edges; weathered. PH: 10 Dims
(ext): 6x7 D (inner): c. 3. SF 9, 110/360,
Structure 4. Fig. 30.11.
Bluegreen translucent wound segmental bead,
one segment and part of a second surviving;
many small bubbles, no weathering. PH: 5
D(max): 3 D(inner): 1. SF 241, 140/295, east of
F236.
Bluegreen, translucent wound segmental bead,
three segments; small bubbles; no weathering.
H: 17 D(max): 5 D(inner): 1. SF 328 145/295,
east of F236. Fig. 30.12.
Bluegreen opaque wound segmental bead, four
segments surviving; small bubbles; no
weathering. H: 14 D(max): 4 D(inner): 1. SF
202, 125/295, east of F236. Fig. 30.13.
Ring bead fragment; bluegreen; opaque; many
bubbles, wound, matt surface. D: 22 H: 9
D(inner): 5. SF 92 110/285, east of Structure 11.
Bluegreen, opaque globular glass bead; small
bubbles; no weathering. H: 4 mm D(max): 4–5
D(inner): 1. SF 283, 145/285, east of F236.
Additional glass beads
Gaming counters
Black, complete; weathered. D: 15 H: 7. SF 112,
60/310, Structure 6.
White, splintered; weathered. D: 14 H: 5. SF
250, pit F213.
Window glass
28
29
30
31
32
33
34
35
36
37
38
39
40
W 18
Matt/glossy
Edge fragment, bluegreen. Dims: 18 x 29 Th: 8.
SF 483, 95/290, Structure 11.
Fragment, bluegreen. Dims: 12 x 26 Th: 4. SF
636, pit F253.
Blown
One fragment. SF 598, 100/365, west of
Structure 4.
One fragment. SF 405, 85/405, east of Structure
1.
One fragment. SF 485, 130/275, east of F236.
One fragment. SF 487, 60/285, Compound 7.
One fragment. SF 488, 60/290, Compound 7.
One fragment. SF 494, 30/290, Compound 7.
One fragment. SF 520, unprovenanced.
One fragment. SF 598, 100/365, area of F134.
One fragment. SF 601, 75/410, area of Structure
1.
One fragment. SF 603, 100/365, area of F134.
One fragment, grozed edge. SF 604, 85/405, east
of Structure 1.
One fragment. SF 622, 40/305, Compound 7.
One fragment. SF 637, 70/310, area of Structure
17.
Lynne Bevan
Four further glass beads were recorded.
43
Opaque, tubular green bead, a type of bead
common in Britain from about the 1st century
onwards (Guido 1978, fig. 37.5). SF 44, 90/380,
area of Structure 2
44 Well-formed, segmented bead of blue-green
glass similar to nos 21–23. F224
45–6 Melon beads made of turquoise faience, one of
which was incomplete. Melon beads are usually
found in 1st and 2nd-century contexts (Guido
1978, 100). SF 23, Area A and unprovenanced.
THE ROMAN POTTERY Annette Hancocks
with contributions by Gillian Braithwaite, Brenda
Dickinson, Kay Hartley, J.M, Mills and Roger
Tomlin
Introduction and methods
A total of 20,267 sherds of pottery was recovered.
The material derived from buildings, pits, ditches and
general overall spreads with an appreciable proportion
coming from hand clearance following machining.
With the exception of the latter material, the pottery
was initially scanned for general spot dating; this was
undertaken by Jane Evans. The pottery from the initial
surface cleaning was not dated but a rapid check for
any new forms was made. All the samian from
whatever source was recorded. A selection of the
pottery from key groups representing good
stratigraphic sequences was recorded in detail. This
amounted to a total of 4065 sherds, 20% of the
stratified material (Table 5). All the pottery
examined was generally in the form of small
abraded sherds.
The key groups were well-stratified, well-dated
groups with 3, 5 and 7 closely associated with
buildings 1, 7 and 11. The pottery was recorded by
context, and fully quantified by count, weight and rim
EVE, using the existing Field Archaeology Unit
Roman pottery recording system. Fabrics, form types
and vessel classes were recorded using the existing
type fabric and form series (Evans 2001), and these
descriptions will not be repeated here. New forms
recognised were integrated into the established form
series and catalogued in greater detail, along with
information concerning production, such as potters’
stamps and wasters, decorative motifs, post-deposition
abrasion, use (sooting and/or residues) and reuse or
FOSSE LANE 1996: SPECIALIST REPORTS
repair (repair holes, lead rivets and counters). Cross
joins and joins were noted where possible. The data
are tabulated to show the fabrics present by period
(Table 6). The pottery is illustrated by key context
group.
The key groups
Group 1
Ditch F134 (contexts 1059, 1064 and
1066); Period 4
A total of 30 sherds representing a minimum of 3
vessels was recovered. Seven fabrics were identified,
with sandy micaceous ware and fine micaceous sandy
ware dominant. The material was generally very
fragmentary and poorly abraded. A single sherd was
recovered from 1066 and three sherds from 1064. No
diagnostic forms were recognised with the exception
of a Black Burnished ware dish (D5.11) of 3rd-century
date from layer 1059, the uppermost fill.
Group 2
Spread F133 (context 1002); Period 2
A total of 293 sherds from a minimum of 36 vessels
was identified. Eleven fabrics were recognised, the
most dominant ones being fine micaceous sandy ware
and samian. All of the material was of 2nd-century
date. With the exception of two unburnt samian sherds
from Les Martres-de-Veyre, the bulk of the samian
was very heavily burnt, a factor which made
identification and close dating of fabrics difficult.
None of the other fabrics identified was burnt. The
samian included Hadrianic-early Antonine forms (Dr
37, Dr 36, Dr 18/31R, Dr 33), as well as later ones
such as Dr 31 and Curle 23, but none of the latest
second century samian forms were present (Fig. 32.6).
The pottery was generally very fragmentary. Two new
forms were illustrated (Fig. 31.1 and 2). These
comprised a Severn Valley ware mortarium stamp (M) and a ‘Belgic’-derived platter (P1.12) in a reduced
fine micaceous sandy ware. Forms previously
recognised include a single Black Burnished ware
bowl (B22.21), a coarse reduced ware bowl and dish
(B23.11 and D1.11), sandy reduced coarsewares
(B16.21, F6.32, J6.21, JC3.42 and JC4.22), a sandy
micaceous coarseware jar (J5.51), a Shepton Mallet
Severn Valley mortarium (M1.12) and a Dressel 20
amphora rim. The majority of the material has a
utilitarian function associated with food preparation
and storage. The only fineware recognised was the
samian, some of which was residual. The pottery may
have been deposited in rubbish used to make up the
ground level.
Group 3
Structure 1; Periods 1 and 2
A total of 878 sherds was recorded dating to the 2nd
century.
Yard F112 (context 1019) - A total of 623 sherds
was recovered with 12 different fabrics identified,
dominated by Black Burnished ware and sandy
reduced coarseware. Forty six vessels were
represented consisting of functional wares such as
Black Burnished ware and reduced greyware cooking
pots, fine sandy greyware beakers, including a face
pot fragment (Fig. 31.3), reduced coarse greyware
dishes (D1.14; Fig. 31.4), occasional Severn Valley
mortaria, and samian cups (Dr 33s) and bowls (Dr
37s). A total of 20 sherds of samian representing 16
vessels was recognised. The material was mainly of
Antonine date, with the latest form being a Walters
79. An earlier Les Martres-de-Veyre vessel was
present.
Shallow pit, F163 (context 1079) - Two fabrics were
recognised, a single oxidised Severn Valley ware
sherd and three sherds of Black Burnished ware, one
of which was a jar type. No samian or other diagnostic
material was recovered.
Pit F117 (context 1018) - A total of 51 sherds was
identified in three fabrics, sandy reduced coarseware,
sandy reduced micaceous ware and sandy oxidised
ware. These are all early fabrics but no diagnostic
dateable forms were observed.
Culvert F120 (context 1016) - Eighty eight sherds
were recorded in eight different fabrics dated to the
later 2nd/early 3rd century. The majority of sherds
comprised oxidised Severn Valley ware, Black
Burnished ware and locally produced coarse and fine
greywares. Only two vessels were represented, a Black
Burnished ware cooking pot form (JC3.13) and a
reduced coarse greyware bowl (B4.21).
Burnt area F143 (context 1022) - A total of 64 sherds
was recorded in four fabrics - oxidised Severn valley
ware, reduced sandy fine and coarsewares and Black
Burnished ware. No diagnostic material was present.
Burnt area F162 (context 1025) - Forty eight sherds
were recovered. A Dr 35 samian sherd of possible
Hadrianic date was present, Six further fabrics were
observed including oxidised Severn Valley ware,
Black Burnished ware, reduced sandy greywares,
sandy micaceous greyware and sandy oxidised wares.
At least five vessels were represented of which one
was a Black Burnished ware dish (D5.11) and another
a new bowl form in a sandy micaceous greyware (Fig.
31.5).
Overall very little diagnostic and dateable material
survives from Structure 1. A date in the late 2nd/early
3rd century seems likely for its abandonment, with
occupation through the 2nd century. The building
itself could well have been used in the preparation
and storage of foodstuffs, given the presence and
dominance of coarsewares over finewares. The
majority of the forms recognised have a utilitarian
function, with vessels such as mortaria, Severn
Valley ware bowls and Black Burnished ware dishes
and storage jars associated with food preparation
and storage.
W19
SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011
TABLE 5: ROMANO-BRITISH POTTERY: FABRIC, SOURCES AND QUANTITIES IN ASSEMBLAGE
Fabric name
Qty
%
Wt (g)
%
Rim EVE
%
BBC
SANDBRF
5ANDMC
SANDRC
SANDRF
SANDROC
SANDRL
SANDRM
SVOXGR
Total reduced
403
356
264
452
357
30
11
165
35
2073
10
9
6
11
9
1
<1
1
1
49
2930
1823
2440
3927
2454
1336
129
2255
300
17594
8
5
7
11
7
4
<1
6
1
49
423
341
141
571
418
27
179
120
2220
4
4
1
6
4
<1
2
1
22
SVOXG
SVOXGM
SVOXGCC(R)
Total Severn Valley
262
42
10
314
6
1
<1
7
1377
384
89
1850
4
1
<1
5
182
27
209
2
<1
2
MISCCR
M1SCCW
SANDOX
SANDOXF
SANDOXG
SANDOXCCW
Totaloxidised
40
2
82
51
4
10
189
1
<1
2
1
<1
<1
5
421
6
566
316
45
37
1391
1
<1
2
1
<1
<1
4
107
20
57
34
218
1
<1
<1
<1
2
SUBTOTAL
2576
61
20835
58
2647
26
BBI
SAVNAK
Total other wares
1197
84
1281
30
2
32
9090
4465
13555
25
12
37
668
10
678
7
<1
7
OXFW
OXFCCW
OXFCCR
Total Oxford wares
3
4
1
8
<1
<1
<1
<1
130
57
1
188
<1
<1
<1
1
22
22
<1
<1
LNVCC
MORTCCW
Total traded wares
2
2
1293
<1
<1
32
1
74
13818
<1
<1
39
700
7
SAMIAN
DRES2O
GIMPORT
IMPORT
Total imported
158
15
1
2
176
4
<1
<1
<1
4
926
1163
93
10
2192
2
3
<1
<1
6
30
21
15
20
86
<1
<1
<1
<1
1
CGW
CREAMI
PM
PREH1STORlC
Total unknown
source
1
1
3
15
20
<1
<1
<1
<1
<1
27
7
11
54
99
<1
<1
<1
<1
<1
11
11
<1
<1
4065
100
36944
100
3444
34
GRAND TOTAL
W 20
FOSSE LANE 1996: SPECIALIST REPORTS
Group 4 Ditches F246 (contexts 2046, 2096, 2097,
2099 and 2100) and F257 (context 2071); Period 1
A total of 288 sherds was recovered from the two ditch
fills suggesting a date in the late 1st or early 2nd
century. Several previously identified forms were
recorded. A total of 53 minimum vessels was observed
in 15 different fabrics. Nine sherds of samian
representing eight vessels were identified. Amongst
the forms recognised was a Dr 37 of Flavian/Trajanic
date, with early 2nd-century material possibly no later
than the mid 2nd century. The wide range of forms
and fabrics recorded was generally of a utilitarian
function.
F246 (2046) - The upper fill of the boundary ditch
contained 84 sherds of 2nd-century date, in 12
different fabrics. Although the sherds are very
fragmentary a range of early forms was recognised.
At least twelve minimum vessels were identified, with
diagnostic rim forms including a Black Burnished
ware dish (D5.12), a reduced sandy fineware bowl
(B17.11), three samian forms (Dr 37, Dr 18/18/31 and
Dr 30). A new flagon form (F9.11) for Fosse Lane, in
a reduced sandy micaceous fabric, was recognised
(Fig. 31.6).
F246 III (2096) - This is the upper fill of the ditch
and comprised 64 sherds, representing 10 different
fabrics and 15 minimum vessels. No new diagnostic
material was recognised, although a variety of
previously published forms confirm the residual
nature of some of the ceramics within the deposit.
Forms identified include Severn Valley ware bowl,
beaker and tankard forms (B10.24, BK2.61 and
T1.12), Black Burnished ware cooking pot forms
(JC3.22, JC3.23 and JC3.31) and lid form (L1.21),
sandy reduced fineware bowls (B23.41 and B20.11),
sandy reduced coarseware dish (D4.11), sandy reduced
micaceous ware wide-mouthed jars (JW3.11 and
JW5.21) and samian Dr 37 and Dr 18/18/31 forms.
F246 III (2097) - A total of 68 sherds were recovered
from the lower fill. The date range for the material
was late 1st/early 2nd century. A total of 12 minimum
vessels was recognised in at least nine fabrics. The
samian was of Flavian/Trajanic date. Diagnostic
material consisted of Black Burnished ware cooking
pot and dish forms (JC3.11, JC4.11 and D5.11), sandy
reduced coarseware lid (L2.11), sandy reduced
fineware bowl, dish and wide-mouthed jar forms
(B13.11, D6.41, JW2.23 and JW5.11).
F246 III (2099) - A single sherd of sandy reduced
fine greyware was recovered from the lowest fill of
this feature. No diagnostic material was recognised.
F246 IV (2100) - A total of four sherds of pottery
was recovered from this fill. Three different fabrics
were identified, although no dating evidence survived.
F257 (2071) - Thirteen fabrics were present in the
single fill of this ditch. A total of 67 sherds was
represented. All the samian was of Hadrianic/Antonine
date. Several diagnostic forms were observed in a
range of fabrics, although no new forms were
recognised. Dateable material comprised a Severn
Valley ware tankard (T1.14), a Black Burnished ware
dish (D6.11), a reduced sandy fineware bowl (B23.31),
reduced coarseware wide-mouthed bowl (JW2.31),
a reduced blue grey, medium to coarse ‘pimply’
micaceous ware beaker (BK3.41), a reduced sandy
micaceous jar (J5.51) and a Dr 35. At least 14 vessels
were recognised.
Group 5
Colluvium layers from Area B (contexts
2030, 2031, 2062); Period 3
A total of 1242 sherds were recovered comprising 177
sherds from 2030, 956 sherds from 2031 and 110
sherds from 2062. An overall total of 43 sherds of
samian representing 32 vessels was recorded.
Layer 2030 - A total of 177 sherds in 14 different
fabrics representing 25 vessels was recognised. A 2ndcentury date was suggested by a Black Burnished ware
bowl (B23.31) and lid (L1.21). Also present were a
sandy reduced fineware bowl (B21.11), a sandy
reduced fineware dish (D6.11), sandy reduced
micaceous ware beaker, jar and wide-mouthed jar
(BK3.12, J10.15 and JW2.22), miscellaneous colour
coat beaker (BK4.22), some Dressel 20 amphorae
sherds and five possible late Iron Age body sherds.
The samian from this layer was Flavian and early 2ndcentury material, again perhaps no later than the midsecond century. At least three Dr 37s were observed
and a Dr 18/18/31.
Layer 2031 - A total of 956 sherds was recovered.
Nineteen different fabrics were present, representing
a minimum of 85 vessels. Severn Valley ware forms
identified include bowl (B21.21), beaker (BK3.75)
and tankard (T2.12) types. In Black Burnished ware
were bowls (B22.21 and B24.12), beaker (BK2.11),
dish (D3.11), cooking pot (JC3.23), wide-mouthed
jar (JW1.11) and lid types (L1.21). In reduced sandy
greyware were a bowl (B21.11), beakers (BK3.41 and
BK3.81) and a lid (L6.11); and in reduced black, fine
sandy ware were bowls (B20.21 and B22.23), a jar
(J14.21) and a wide-mouthed jar (JW2.21). Sandy
reduced coarseware forms were a bowl (B22.23), a
beaker (BK3.21), a flagon (F5.11), jars (J14.31 and
J8.21), a cooking pot (JC3.41) and wide-mouthed jars
(JW2.21 and JW3.11). Also present were a sandy
reduced micaceous ware bowl (B22.22) and reduced
colour coat beakers (BK3.12 and BK3.65). The
samian was exclusively Antonine in date with the
latest vessels (joins 2094) dated AD 150–190, a Dr
37 bowl (Fig. 32.4). Other samian forms identified
included Dr 31 and Dr 18/31 or 31 types. Some
Dressel 20 amphorae sherds were present.
Two 4th-century coins were found which must be
viewed as intrusive.
Layer 2062 - A much smaller quantity of pottery
W 21
Fabric name
Qty
P2
Wt
%
(g) rim
Av
wt
Qty
P3
Wt %
(g) rim
Av
wt
Qty
SANDRF
5ANDBRF
5ANDRM
SANDMC
SANDRC
BBC
SVOXGR
SANDRL
SANDRGC
SVOXG
SVOXGM
SANDDXF
SANDOX
5ANDOXCCW
SANDOXG
SVOXOCC(R)
MISCCW
MISCCR
Total regional
101 618
129 513
14 383
107 1100
157 1305
124 665
22 145
1
9
5
119
116 424
35 345
12
66
3
11
3
6
1
3
830 5712
191
76
46
58
195
163
28
13
7
830
6
4
27
10
8
5
7
9
24
4
10
6
4
2
3
7
1
2
8
41
48
13
8
4
27
4
39
1
196
35
61 15
140 25
454 48
590 36
356 56
61
56
364 42
12
90
8
2227 222
35
30
18
11
12
27
8
14
13
3
2
8
11
BBI
SAVNAK
OXFW
OXFCCR
OXFCCW
LNVCC
MORTCCW
Total traded
412 2328
13 444
2
1
2
74
429 2847
165
165
6
34
<1
37
7
69
57
126
563
3542
4105
80
10
90
8
62
33
648
13
1
4
666
PM
CGW
CREAM
PREHISTORIC
Total unknown source
SAMSG
SAMCG
DR20
Total imported
TOTAL POTTERY
P4
Wt
(g)
246 1754
197 1139
131 1616
99
731
225 1824
232 1619
1
15
5
47
25 1217
115
568
2
22
33
224
38
459
9
29
4
45
7
83
2
6
39
418
1410 11816
P5
% Av
rim wt
Qty
Wt
(g)
%
rim
Av
wt
227
210
108
35
301
145
27
102
14
34
50
20
107
1380
7
6
12
7
8
7
15
9
49
5
11
7
12
3
11
12
3
11
8
9
28
12
17
22
34
4
1
4
1
6
2
140
47
110
116
155
208
290
79
17
21
5
26
6
1080
40
39
59
67
10
215
5
4
10
9
9
8
20
17
5
5
4
3
8
5716
473
82
57
6328
366
15
381
9
36
82
14
10
67
1
2
1
71
478
6
48
I
533
57
7
64
7
6
24
1
8
-
-
-
-
-
-
-
-
3
1
15
19
11
7
54
72
11
11
4
7
4
4
1
1
27
27
-
27
27
4
87
1
92
14
629
131
774
10
39
21
70
4
7
131
8
1
3
4
3
39
42
-
3
13
10
5
54
14
73
8
229
1032
1269
10
10
2
15
74
17
1
1
2
1
3
4
5
5
1
3
2
1351 9333 1065
7
326
6374 312
20
2168 19485
1782
9
214
1644
284
8
SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011
W 22
TABLE 6: SUMMARY OF FABRICS BY PERIOD (KEY GROUPS ONLY)
FOSSE LANE 1996: SPECIALIST REPORTS
was recovered from this layer. Twelve fabrics were
identified representing a total of 110 sherds. At least
16 minimum vessels were recognised. A single new
diagnostic rim form was identified from the group, a
reduced black, fine sandy ware lid (L4.12, Fig. 31.7).
The only other dateable piece recovered was a Black
Burnished ware cooking pot (JC3.23). The samian
comprised four sherds of Dr 31, probably exclusively
Antonine in date.
As with layer 2031 the pottery evidence was
contradicted by the presence of three 4th-century
coins.
Leaving aside the coin evidence the pottery suggest
a 2nd-century date for the three colluvium layers.
Group 6
Period 2
Oven F249 (2053, 2091, 2093, 2094);
A total of 53 sherds was recovered from the fill (2053)
and rakeout (2093) of the oven, with a further 180
sherds recorded from the yard or building floor surface
2094 associated with F249 and from layer 2091
nearby. Seven sherds came from 2053. A single vessel
was recognised in a Black Burnished ware bowl form
(B22.21). The only other fabric recovered was a
Severn Valley ware sherd. Forty six sherds came from
layer 2093 and comprised six fabrics, including Black
Burnished ware (D5.12), reduced black, fine sandy
ware, sandy reduced coarsewares, sandy micaceous
wares, sandy oxidised finewares and some prehistoric
material. The material was dateable to the 3rd century.
At least four vessels were present.
The pottery from layer 2094 comprised 8 different
fabrics and ten minimum vessels. Severn Valley ware
forms included bowl and tankard forms (B9.13 and
T2.12). Black Burnished ware was in jar forms (J1.21
and JC3.23), and reduced fine micaceous sandy ware
in a bowl form (B20.14). A Dr 37 form dated AD
150–190 was present (Fig. 32.4), and five sherds from
three Hadrianic and Antonine vessels including 3
sherds joining with 2031.
The pottery from the stone rubble platform to the
east of F228 (2091) was very heavily abraded and
fragmentary in nature. A total of 10 different fabrics
was identified and at least 11 vessels. Very little
diagnostic material was recovered, although samian
forms Dr 33 and Dr 30 were observed, both of
Hadrianic/Antonine date.
Group 7
5
Structure 7 (2024, 2025, 2029); Period
Three contexts were associated with this structure. A
total of 417 sherds was recovered, representing a
minimum of 37 vessels.
From the floor within Structure 7 (2024), a total of
33 sherds, representing at least 1 vessel, was recovered
and dated to the 3rd century. Generally the material
was poorly abraded and small in size. Only six fabrics
were identified, none providing any diagnostic
material, although Black Burnished ware and sandy
oxidised wares were present.
Layer 2025 represented a rubble deposit in Structure
7. A total of 208 sherds was represented in 15 different
fabrics and making 25 minimum vessels, dating to
the late 3rd/4th century. Much residual second-century
material was recognised, including Savernake ware,
Severn Valley ware and samian. No new forms were
observed. Diagnostic wares present included Black
Burnished ware bowls, dishes and cooking pot forms
(B24.11, B24.13, D5.11, JC4.11 and JC3.23); coarse
greyware copies of BB1 forms such as dish (D5.11)
and cooking pot forms (JC3.23, JC4.11 and JC4.21);
sandy reduced coarseware beakers, jars and lids
(BK3.81, J9.31 and L3.11); Oxfordshire fine white
mortarium (M2.71) and four very small sherds of
samian of mixed date.
Layer 2029 represented burning in Structure 7. A
total of 176 sherds was identified, in 11 different
fabrics and 11 minimum vessels. The layer comprised
mainly material dating to the early 3rd century. Little
diagnostic material was recovered, with the exception
of a Black Burnished ware dish (D5.11) and cooking
pot (JC3.31), and a coarse reduced sandy greyware
dish copy (D5.11). A much later date was suggested
by three 4th-century coins.
Group 8
Midden (2092); Period 2
The layer contained 159 sherds in ten different fabrics
of late 1st/early 2nd century date with a minimum 22
vessels present. Diagnostic forms include Severn
Valley ware bowls (B3.11, B7.11 and B9.13), Black
Burnished ware dish and cooking pot forms (D1.11
and JC3.23 and JC3.31), reduced coarse greyware
BB1 copies of cooking pots (JC3.13, JC3.23 and
JC3.31), a reduced fine micaceous sandy ware tankard
(T1.11), sandy reduced coarseware beaker and jar
(BK3.21 and J6.21), a reduced blue grey medium
sandy ware bowl and flagon (B22.22 and F5.11),
brown/buff to grey, medium to coarse ‘pimply’
micaceous ware bowls and a wide-mouthed jar
(B13.11, B22.22 and JW2.22), and samian comprising
two small sherds and two stamped Dr 33 bases (stamps
Fig. 33.1 and 33.8) of mixed dates from first through
the second centuries.
Group 9
Pits F251 (2060) and F256 (2056, 2057,
2070 and 2073), Period 4; and hearth F250 (2054),
Period 2
The fill of F251 contained over 264 sherds of pottery
from a minimum of 21 vessels in 11 different fabrics.
The assemblage dated to the 3rd century and included
diagnostic material in the form of Black Burnished
ware bowls (B23.11 and B23.31), a dish (D5.11), a
W 23
SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011
jar (J10.41) and cooking pots (JC3.24 and JC3.31).
Other forms recognised include sandy reduced
coarseware Beaker forms (BK3.21 and BK3.81) and
three small sherds of samian including fragments of
Dr 37 and Dr 31 forms of Hadrianic-Antonine date.
Two 4th-century coins demonstrated that all the
pottery was residual.
A total of 116 sherds was recovered from F256.
The upper fill, 2056, contained 24 sherds
representing at least two vessels with eight 8
different fabrics. No new forms were observed, with
the only diagnostic material comprising two Black
Burnished ware cooking pot forms of 3rd-century
date (JC3.23 and JC4.21). In addition, a single
sherd of samian (Dr 43/Curle 21) of late 2ndcentury date was present.
Layer 2057 contained 31 sherds in 10 fabrics
representing at least nine vessels. Black Burnished
was the most common fabric represented. Diagnostic
material consisted of several forms including a dish
(D5.11), bowl (B23.11) and cooking pots (JC3.23 and
JC3.31). Additionally a single sherd of prehistoric
pottery was identified.
Layer 2070 contained 38 sherds in 6 fabrics
representing at least five vessels. The material was of
late 2nd/early 3rd date. Very little dateable material
survived, with the exception of a Black Burnished
ware bowl (B23.11) and a sandy reduced fineware
jar (J10.32). A good dateable piece of samian was a
Dr 37 bowl fragment (Fig. 32.3) dating to AD 150–
80 and a Dr 18/31R or 31R.
The lowest fill (2073) provided very little
dateable material from 23 sherds. A single minimum
vessel was represented, although six different
fabrics were observed. The only diagnostic form
was a Black Burnished ware dish (D5.11) of 3rdcentury date.
A total of 13 sherds was recovered from F250. The
pottery consisted of small and very abraded sherds.
Very little dateable material was observed, although
two fragments of Dr 18/31 and Dr 30/37 samian of
Hadrianic-Antonine date formed part of the
assemblage.
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Illustrated sherds (Fig. 31)
24
1
2
3
4
5
W 24
Mortarium (M); see comments by K Hartley; MSVOXG, 8%. Layer 1002, F133, Period 2, KG2
Platter with Gallo-Belgic derivation; P1.12
SANDRF, 9%. Layer 1002, F133, Period 2,
KG2
Decorated sherd, facepot fragment; SANDRF.
Layer 1019, Period 2, KG3
Dish with flanged rim; D1.14 SANDRC, 31%.
Layer 1019, Period 2, KG3
BB1-type bowl; conical bowl with flat grooved
and flanged rims; B23.42 SANDMC, 7%. Layer
1025, F162, Period 2, KG3
25
26
Flagon with bifurcated rim; F9.11 SANDMC,
10%. Layer 2046, F246, Period 2, KG4
Lid with internal groove; L4.12 SANDBRF,
11%. Layer 2062, Period 4, KG5
Bag-shaped beaker; BK3.42 SVOXGR, 16%.
Layer 3003, F301, Period 5
Bag-shaped beaker; BK3.44 SVOXGR, 30%.
Layer 3003, F301, Period 5
Bag-shaped beaker; BK3.45 SVOXGR, 21%.
Layer 3003, F301, Period 5
Bowl with rim overhanging internally flanged
bowl (segmental); B2.22 SANDBRF, 25%.
Layer 1021, F142, Period 2
Bowl with rim overhanging internally flanged
bowl (segmental); B23.43 BBC, 19%. Layer
2037, F211, Period 3
Dish, for graffito see Tomlin below; DSANDBRF, 7%. Layer 2048, F217, Period 4
Dish with plain rim and increasingly splayed rim;
D4.12 SVOXG, 16%. Layer 1001, Period 4
Jar with near triangular rim, ?waster; J6.21
SANDRC. Layer 3003, F301, Period 5
Globular jar with inturned rim and external lid
seating; J16.12 SANDRM, 12%. Layer 3000,
F209, Period 2
Jar with splayed rim of near equal girth, obtuse
cross-hatch burnish surmounted by a groove; for
graffito see Tomlin below, JC4.11 BB1, 19%.
Layer 2019, F225, Period 5
Mortarium; M5.31 GIMPORT, 15%; Antonine
(Hartley 1991, fig. 84, C51). Layer 1037, F150,
Period 6
Platter, Gallo-Belgic derivation; P3.11
SANDBRF, 10%. Layer 1049, F134, Period 4
Platter, Gallo-Belgic derivation; P4.11
SANDBRF, 10%. Layer 2017, Period 2
Platter, Gallo-Belgic derivation; P5.11
SANDBRF, 6%. Layer 2017, Period 2
Tankard; T4.11 IMPORT. Layer 2020, F223,
Period 6
Decorated body sherd, facepot fragment; Gillian
Braithwaite comments that this type is not easily
paralleled by West Country military examples.
It is very late in date. SVOXGR. Layer 3003,
F301, Period 5
Decorated body sherd. Layer 2082, F260, Period
4
Decorated body sherd; (T-) IMPORT. Layer
2004, F213, Period 7
Small cup; MS7.11 SANDOX, 7%. Layer 2017,
Period 2
Mortarium Kay Hartley
Two joining sherds from the bead and upper part of
the flange of a mortarium (Fig. 31.1). The bead and
upper surface of the flange are blackish grey, almost
FOSSE LANE 1996: SPECIALIST REPORTS
Fig. 31 Roman pottery
certainly the result of overfiring, though the vessel
may have been a second rather than a waster. The
inside fabric and inner surface are ‘cocoa’ brown
(Munsell 10RS/4), and slightly rough to the touch.
No obvious slip survives but the brown stains in and
near the stamp may be the remains of a discoloured
red-brown slip. The fabric is fine-textured and the
moderate, tiny and small quartz and slag inclusions
are barely visible at less than x20 magnification. Two
quartz trituration grits survive.
The left-facing stamp is fragmentary, but enough
survives to show that it is from an unrecorded die;
other stamps from the same die should be identifiable.
The large bead, the fabric, the type of stamp and the
provenance leaves no doubt that it is a product of the
Shepton Mallet pottery workshops, which were active
within the period AD 100–140, perhaps early in the
century.
Samian: decorated wares J.M. Mills
Sherds representing some 86 or 87 vessels are
described in the catalogue below. There will inevitably
have been a few rim and base sherds within the
remaining unrecorded material, so that the total
number of decorated vessels is probably slightly
higher. The decorated wares date from the Neronian
period through to the late Antonine. Although there
are no apparent gaps in the supply of these vessel
types only three of Hadrianic date were recorded from
a total of 65 closely dated vessels. Many of the sherds
are small and with fragments of decoration surviving.
Of the 28 decorated vessels from Southern Gaul,
26 are probably from La Graufesenque. One is a late
1st or early 2nd-century Dr 37 from Montans (Fig.
32.1), the source of the other, a Neronian Dr 29, was
not identified. The decorated wares include seven Dr
29s (SG), one Dr 29 or 37 (SG); four Dr 30s (SG 1,
W 25
SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011
CG 3); 67 or 68 Dr 37s (SG, 18; Les Martres 15, CG
Lezoux 35 or 36), one Dr 67 (SG) and one Dr 64
(CG). The latter, although not unusual, is a less
common form (Fig. 32.6) and is probably the work of
Libertus. The discrepancy in the numbers of Central
Gaulish Dr 37s results from the possibility that two
of the sherds may be from the same vessel.
14
15
16
17
Catalogue
18
The catalogue is presented in approximate date of
manufacture order.
19
1
20
2
3
4
5
6
7
8
9
10
11
12
13
W 26
Dr 29, SG. Base with fragment only of lower
zone of straight gadroons; AD 55–70. 120/260
Dr 29, SG*. Small body sherd with two small
leaves from a scroll design extant. The fabric is
almost certainly not from La Graufesenque*
although its source is not identified; Neronian.
80/295
Dr 29, SG. Edge of lower zone of decoration
fan-shaped plant only remains. Similar plants
occur on bowls from La Graufesenque stamped
by Germanus (his earlier work, not using his own
moulds); Late Neronian and early Flavian type;
AD 60–75. 120/260
Dr 29, SG. Carination of bowl with winding
scroll in upper zone and zonal decoration below
with inhabited medallions. Small spirals used as
infills in both upper and lower designs; AD 60–
75. 130/340
Dr 29, SG. Foot ring fragment; Neronian or early
Flavian. Layer 3011
Dr 29, SG. Fragment of plain band and bead rows
from carination; AD 70–85. Layer 2011
Dr 37, SG. Basal wreath only; AD70–90. 75/285
Dr 37, SG. Body sherd in the style of Germanus
with ovolo with corded tongue with rosette
terminal; no border below, with only fragments
of decoration extant; AD 70–90. Layer 2030
Dr 37, SG. Lower part of decoration with a
wreath of S-shaped gadroons with wavy line
below and a basal wreath of bifid leaves; AD
70–90. 125/265
Dr 29, SG. Fragment of plain band with large
bead row either side and a scrap of ?winding
scroll; Early-mid Flavian. 75/280
Dr 37, SG. Body sherd with fragments of two
double-bordered medallions or festoons and a
long, triangular leaf; AD 75–95. Layer 2048
Dr 37, SG. Ovolo with trident tongue and wavy
line below. M Crestio regularly used a version
where the trident is blurred as in this example; c
AD 75–100. 80/270
Dr 37, SG. Body sherd with fragment of zoned
decoration comprising panel infilled with
diagonal wavy lines to leave triangular zone
inhabited by a lion running to right (similar to
21
22
23
24
25
26
27
28
29
30
31
32
O.1400); Flavian. Layer 3007
Dr 37, SG. Fragment of decoration, too small to
identify; Flavian. 50/300
Dr 67, SG. Fragment from top of moulded
decoration; Flavian. 100/375
Dr 30, SG. Fragment of decoration, too small to
identify; Flavian. 35/320
Dr 37, SG. Fragment of medallion only; Flavian.
Layer 2030
Dr 37, SG. Body sherd with fragment of spiral
with central rosette; Flavian. 75/280
Dr 37, SG. Fragment of lower edge of decorated
zone; Flavian. Layer 3011
Dr 37, SG (overfired). Fragment of decoration
with wavy line border; Flavian. 90/400
Dr 37, SG. Small body sherd with fragment of
trifid-tongued ovolo; AD 85–110. 105/386 (layer
1001)
Dr 37, ?SG. Ovolo with trifid tongue, similar if
not identical to ovolo of Mecator. Plain, single
bordered medallion below contains a trilobed
motif which appears on an unpublished stamped
bowl from Leicester (Brenda Dickinson pers.
comm.); AD 85–110. Layer 3007
Dr 37, SG. Trident-tongued ovolo with bead row
below. Fragment of animal body below; FlavianTrajanic. Layer 2096
Dr 37, SG. Trident-tongued ovolo; FlavianTrajanic. Layer 2017
Dr 37, SG. Blurred ovolo only; Flavian-Trajanic.
55/275
Dr 37, SG. Fragment of moulded decoration;
Flavian or Trajanic. u/s
Dr 29 or 37, SG. Body sherd with horizontal bead
row fragment; 1st century. Layer 1040
Dr 37, SG (Montans) (Fig. 32.1). Body sherd
with dog (with collar) running left, similar to
motifs used in Central Gaul eg O.1980,
O.1989A. Bowls made at Montans sometimes
have a similar dog running to the right. The fabric
is very pale with occasional large (<1mm) red/
brown ?iron inclusions and is not a fabric used
by the later Montans potters and so is unlikely
to be later than the very early second century
(Brenda Dickinson pers. comm.). Another sherd
of similar fabric was recovered from 80/395 and
may derive from the same vessel; Trajanic. 75/395
Dr 37, CG (Les Martres-de-Veyre). Fragment of
ovolo; Trajanic. Layer 1019
Dr 37, CG (Les Martres-de-Veyre). Rim sherd
with ovolo Rogers B44 with wavy line (Rogers
A24) below. Potter of the Rosette; AD 100–120.
Layer 2030
Dr 37, CG (Les Martres-de-Veyre). Body sherd
with ovolo Rogers B28. Probably Potter X-2 or
X-3; AD 100–120. 105/360
Dr 37, CG (Les Martres-de-Veyre). Body sherd
with fragment of ovolo Rogers B37 with wavy
FOSSE LANE 1996: SPECIALIST REPORTS
Fig. 32 Decorated samian
33
34
35
36
line below. Fragment of rosette Rogers C280
and fragment of ?Perseus (O.234); Potter X-4
(Igocatus); AD 100–120. 105/360
Dr 37, CG (Les Martres-de-Veyre). Rim with
ovolo Rogers B29; Potter X-4 (Igocatus); AD
100–120. 115/275
Dr 37, CG (Les Martres-de-Veyre). Body sherd,
ovolo double impressed and not identifiable,
wavy line below with tail of seahorse only
surviving. The seahorse is probably that used by
Potter X-12 (Stanfield and Simpson 1990, pl 40,
462); AD 100–120. 125/260
Dr 37, CG (Les Martres-de-Veyre). Fragment
with rosette Rogers C292 and wavy line above
(or below); Potter X-12; AD 100–120. 75/275
Dr 37, CG (Les Martres-de-Veyre). Fragment
with horizontal bead row and festoon Rogers F8.
37
38
39
40
41
Probably Potter X-13 or X-14. AD 100–120. 90/
270
Dr 37, CG (Les Martres-de-Veyre). Small body
sherd, extant decoration comprises the junction
of Rogers motif F74 with another, possibly J169,
rather than a festoon or medallion with leafy
border. The base of a ?trifid leaf sits at the
junction; AD 100–120. Layer 2076
Dr 37, CG (Les Martres-de-Veyre) (Fig. 32.2).
Fragment of ovolo with bead row below; ovolo
overlaps; AD 100–120. Layer 2017
Dr 37, CG (Les Martres-de-Veyre). Small sherd
with blurred ovolo; AD 100–120. Layer 2043
Dr 37, CG (Les Martres-de-Veyre). Chip with
fragment of blurred ovolo; AD 100–120. Layer
2012
Dr 37, CG (?Les Martres-de-Veyre). Scrap of
W 27
SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011
42
43
44
45
46
W 28
moulded decoration; ?AD 100–120. 110/285
Dr 37, CG (Les Martres-de-Veyre) (Fig. 32.3).
Seven sherds from the same vessel, not all
joining. The ovolo is that of Libertus (Rogers
B214) with a poorly executed bead row below.
The bowl is sloppily finished as is often the case
at Les Martres-de-Veyre. The bowl is decorated
with Ulysses in his ship, a variant of O.982A;
sherds with fragments of three of these were
recovered. The decoration is infilled with a
variety of nautical motifs including turtles
(O.2156 variant), fish (O.2417), and tritons or
sea horses (not in Oswald). A bead row towards
the lower edge of the decoration represents the
sea, below which, on the sherd where this
element of the design survives, swim fish. This
same combination of motifs can be seen on an
unstamped bowl from London, although
seahorses replace fish in the lower zone on that
bowl (Stanfield and Simpson 1990, pl 53, 626);
AD 115–130. Layer 2017 and 60/275
Dr 37, CG (Les Martres-de-Veyre) (Fig. 33.3).
See also stamps below. Body sherd with seahorse
facing right similar to that on an unstamped bowl
from London (Stanfield and Simpson 1990, pl
53, 626). Fragment of a possible stamp ]O or O[
at edge of sherd (stamp cat. no. 3). The slip is
damaged and the sherd is of incomplete thickness
and worn, however, the curvature of the sherd,
the depth and detail of the moulding as well as
the fabric suggest that this may be a fragment of
the Libertus bowl (see 2017 etc) above; ?AD
115–30. 75/285
Dr 37, CG. Lowest part of panelled decoration
of large double-bordered medallions with
vertical bead rows between, these having
terminal rings and flanked by astragalus Rogers
R12. The two figures within the medallions are
Abundance O.802 and Bacchus O.571. To the
right is caryatid 0.1199. The work of Divixtus;
all these elements can be seen on stamped bowls
of his (Stanfield and Simpson 1990, pl 115, 3
and pl 116, 8) from Silchester and Coventry, and
Carlisle and Leicester respectively; AD 140–175.
u/s 4
Dr 37, CG (KG9; Fig. 32.4). Body sherd with
ovolo Rogers B135 and corded border A36
below. The extant decoration comprises a plain
festoon with ‘seahorse’ O.52A. In the style of
Ivstus; AD 150–180. Layer 2070
Dr 37, CG. Base and lower part of decoration
only comprising vertical borders Rogers A10,
stands Rogers Q84, and the legs of figure O.569
but without the cup. Possibly the work of Illixo,
who used the figure (with the cup) and the
borders. The bowl retains a lead rivet from a
repair to the base. The lowest part of the interior
is heavily worn with no slip remaining; AD 150–
47
48
49
50
51
52
53
54
55
180. 55/285
Dr 37, CG. Fragment with ovolo Rogers B145
used by Cinnamus, Carantinus and Illixio; AD
150–180. u/s 3
Dr 37, CG. Body sherd, heavily burnt. Ovolo
fragment probably Rogers B143, with bead row
below. Figure within plain double bordered
medallion is dancer O.348 with broken hand;
Cinnamus; AD 150–180. 85/285
Dr 37, CG. Body sherd with fragment of panelled
decoration with vertical bead row and figure of
Venus at an altar O.322 to the right. Probably
Cinnamus; AD 150–180. Layer 1001
Dr 37, CG. Body sherd from bowl in the style of
Albucius. With dog running to left (possibly a
small version of O.1984) and leaf tips in the
ground, Rogers J145, but the edge only; AD
150–180. 125/265
Dr 37, CG (KG 5 and 6; Fig. 32.5). Two groups
of joining sherds from the same vessel with ovolo
Rogers B105 with blurred bead row below. The
bowl is decorated with a freestyle design of wild
animals, dogs and horses, with and without
riders. The figures used include a variant of
horseman O.246, horse O.1911, and bear
O.1589. Above the bear are two very poorly
impressed leaf-tip fillers. This could be the work
of either Paternaus or Albacus, however, the leaftips may be those of Albacus (compare with
bowls from Corbridge and London – Stanfield
and Simpson 1990, pl 123.33, 35, 38, 41 and
42); AD 150–190. Layers 2031 and 2094
Dr 37, CG. Body sherd with ovolo Rogers B20;
Secundus i; AD 125–140. 110/270
Dr 64, CG (Fig. 32.6). Rim fragment with two
parallel grooves at top of decoration. Of the three
figures two are complete and identifiable, cupid
O.422 and kneeling stag O.1752; both were used
by Butrio. The discus (Stanfield and Simpson
1990, fig. 12.4) used by Libertus appears below
and to each side of the cherub. The stag and
discus can be seen together on an unstamped
form 64 beaker from London (Stanfield and
Simpson 1990, fig. 52, 613) which is attributed
to Libertus. There is no evidence for Butrio using
the discus, suggesting that this may be the work
of Libertus; Hadrianic, 80/395
Dr 37, CG (Fig. 33.2). Stamp in plain area below
decoration (stamp no 2); AD 125–150. 60/265
Dr 37, CG. A bowl with the distinctive buds of
the Cerialis ii – Cinnamus ii group (Rogers
J178), here in very clear impressions and so
probably indicating an early product of the
potter. Other figures include Venus (O.286),
winged centaur (O.735A); bear (O.1609) with
head squashed (?by a thumb). The striated
divider and row of circles below both used by
this group; AD 135–150. 115/380
FOSSE LANE 1996: SPECIALIST REPORTS
56
57
58
59
60
61
62
63
64
65
66
Dr 37, CG. Rim sherd with ovolo Rogers B108
with wavy line A23 below. Maccius and
Maccarirra used the ovolo; AD 130–160. Layer
2043
Dr 37, CG. Body sherd with ovolo Rogers B47
(or a variant of it) and a small rosette Rogers
C120. The lack of a border below the ovolo is
paralleled at Lincoln on a bowl in the style of
Tetturo (Brenda Dickinson pers. comm.); AD
130–160. 85/270
Dr 37, CG. Body sherd with ovolo B144 with
bead row below. Cerialis ii/Cinnamus ii; AD
135–170. Layer 2068
Dr 37 CG. Joining body sherds with ovolo
Rogers B144 with bead row below. Moulding
poor. Extant decoration comprises figure in
arcade between vertical bead rows; Cinnamus/
Cerialis group; AD 135–170. 60/285
Dr 37, CG. Joining body sherds with ovolo
Rogers B144, with small bead row below. Part
only of panelled design extant comprising
vertical bead row, open ring and fragment of
motif O.234 (Perseus). Cerialis ii/Cinnamus ii
group; AD 135–170. 105/365
Dr 37, CG. Body sherd with upper part of
decoration comprising ovolo Rogers B107 with
bead row below. Decoration is panelled with
only the head of Pan O.709 extant. Paternus; AD
150–190. 50/405
Dr 30, CG. The decoration is panelled with bead
rows dividing. The surviving elements include
shell Rogers U76 in a plain double-bordered
medallion and a column including vase? Rogers
T16, Q59 and astralagus R18. Almost certainly
the work of Doeccus; AD 160–190. Layer 2050
Dr 37, CG. Lower part of design with single line
delimiting the decorated zone. The extant
decoration comprises the lower legs and feet,
probably of Venus O.278 and, to the left,
fragments of a winding scroll of two plain lines
terminating with acorn Rogers U87. Both the
Venus and the acorn appear on sherds from
Corbridge (Stanfield and Simpson 1990, pls
125.8, 126.16 respectively). The Venus has a
damaged left foot as, from the illustration, the
Corbridge example appears to have. The scroll
is similar to the Corbridge example and may be
seen, although in reverse, on a sherd from
Dorchester (Pengelly 1982, fig. 33, 38); Iullinus
style; c AD 160–190. Layer 2066
Dr 37, CG. Body sherd with ovolo (Stanfield and
Simpson 1990, fig. 44.2), rosette Rogers C274
and beaded ring E58; Style of Doeccus; AD 160–
190. 75/280
Dr 37, CG. Scrap of decoration, rosette Rogers
C144. Probably Paternus; AD 160–195 (Brenda
Dickinson pers. comm.). 55/275
Dr 37, CG. Chip with fragment of rosette-
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
tongued ovolo and wavy line below; Hadrianicearly Antonine. 75/275
Dr 37, CG (burnt). Incomplete ovolo with corded
tongue; Hadrianic or early Antonine. Layer 2021
Dr 37, CG (heavily burnt) KG 2 (Fig. 32.7). A
substantial portion of a panelled bowl with ovolo
Rogers B17 with large bead row (?A2) below,
and panels divided with bead rows. The panels
alternate between large human figures and a
festoon inhabited by a cockerel with another
motif (missing) below. Figures include Bacchus
O.566, Venus O.278, Cock O.2348 (or variation)
and a small figure similar to O.658B. Attribution
is difficult, the ovolo is probably that of Paternus
iv (Brenda Dickinson pers. comm.), however his
repertoire of figures is not yet known; probably
Hadrianic or early Antonine. Layer 1002
Dr 37, CG. Two sherds, not joining but likely to
be from the same vessel. Fragment of panelled
decoration with vertical bead row and small
fragment of a human figure (unidentified);
Hadrianic or Antonine. Layer 1019
Dr 37, CG. Body sherd with fragment of
inhabited scroll, motif too fragmentary to
identify; Hadrianic-Antonine. Layer 2003
Dr 30 or 37 CG. Scrap with decoration;
Hadrianic or Antonine. 50/300
Dr 37, CG. Body sherd with small fragment of
ovolo; Hadrianic or Antonine. 70/410
Dr 30 or 37, CG. Decorated fragment. Hadrianic
or Antonine. 75/405
Dr 30 or 37, CG. Decorated fragment (ext surface
only); Hadrianic or Antonine. 75/405
Dr 37, CG. Fragment of lower edge of decorated
zone; Antonine. Layer 2031
Dr 37, CG. Fragment of decoration; Antonine.
Layer 2031
Dr 30, CG. Body sherd with fragment of stag
running to left with the antler tines pointing
upwards. No exact parallel in Oswald; Antonine.
Layer 2066
Dr 30 or 37 CG. Body sherd with vertical bead
row and ?foot of human figure to left; Antonine.
Layer 1040
Dr 37, CG. Body sherd with ovolo ?B17;
Antonine. 55/405
Dr 37, CG. Body sherd with most of external
surface missing. No identifiable decoration
survives; Antonine. 55/405
Dr 37, CG. Fragment only, with single bordered
medallion; Antonine. 115/260
?Dr 30, CG. Two chips, probably Dr 30;
Antonine. 50/315
Dr 37, CG. Lower part of decoration with vertical
bead row with ring at terminal; standing Bacchus
to the left (O.566); Antonine. 70/400
Dr 37, CG. Fragment of panelled decoration with
vertical bead row and part of figure, possibly
W 29
SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011
85
86
87
Venus (similar to O.279); Antonine. 75/400
Dr 37, CG. Fragment of panelled decoration with
vertical bead row and leg of standing figure,
probably Perseus (O.234) to the right; Antonine.
90/385
Dr 37, CG. Chip with fragment of decoration;
Antonine. 90/385
Dr 37, CG. Body sherd with upper part of
decoration. There is no ovolo on this bowl. The
only figure is a cupid O.440; Mid–late Antonine.
75/405
4
5
Potters’ stamps Brenda Dickinson
Each entry gives: excavation number, potter (i, ii,
where homonyms are involved), die, form, reading,
published example (if any), pottery of origin, date.
Superscript (a), (b) and (c) indicate:
(a) A stamp attested at the pottery in question.
(b) Not attested at the pottery in question, but other
stamps of the potter known
from there.
(c) Assigned to the pottery on the evidence of
fabric, distribution, etc.
Ligatured letters are underlined
6
7
Catalogue (Fig. 33)
1
2
3
W 30
Dagomarus 3b 33a [DAGOMA]RVSFV (final
letter uncertain) Les Martres-de-Veyre a. .
Dagomarus worked at Les Martres under Trajan
and at Lezoux in the Hadrianic period. There
are many vessels with this stamp in Les Martres
fabric in the London Second Fire groups. c AD
100–120. Layer 2092, KG8
Gratus ii la 37 (stamped upside down below the
decoration, after moulding) [GR]ATI.M
Lezouxb. This appears on a bowl from York with
a mould-stamp of Quintilianus i (Stanfield and
Simpson. 1990, pl. 68, 7). It occurs also in the
Rhineland, where the import of Central Gaulish
ware seems to have ceased around the middle
of the second century, and at South Shields. c.
AD 125–150. 60/265
Libertus ii 6a 37 O[FLIBERTI] (Dechelette
1904, 282, no 110, 39). The presence of Libertus
wasters at Lezoux is evidence that he worked
there, but this bowl is in one of the fabrics in the
range produced at Les Martres-de-Veyre, and so
will have been from a mould traded there, or
brought from Lezoux. Dating evidence for this
potter is slight, but his limited distribution and
his use of non-standard fabrics suggest that he
was rather earlier than the main body of
Hadrianic potters working at Lezoux. The
similarities between his style of decoration and
that of Butrio suggest Trajanic–Hadrianic date.
8
9
10
Layer 2017
Mettius 1a’ 18/31R METTI.MA Lezouxc. This
is from a broken die which originally had ends
to the frame. It occurs at Mumrills and
Newstead, while a stamp from a different die is
known from the Saalburg Erdkastell (before AD
139). c AD135–160. Layer 2072
Osbimanus-Cadgatis 2a 33 OSBIMLCL
Lezouxa. This stamp records a partnership, or
joint venture, of two potters, whose names are
extrapolated from a stamp on which they are
more complete. It was used only on cups, mainly
of form 33, but with one example on form 46.
Both potters worked independently at some stage
in their careers. Their separate ranges of forms,
which include 18/31R, 31R, 79 and Ludowici
Tx, suggest activity c AD 150–180. 50/390
Pateratus 5a PATIIRATV Lezouxb. This stamp
is known from Birdoswald, Chesterholm and
Corbridge, and has been noted once on form 18/
31R. Pateratus’s use of other dies on forms 18/
31, 27 and 81, combined with the site evidence,
suggests a range c AD 135–165; see also graffito.
85/300
Paterclinus 4a 18/31R or 31R [PLTER]CLINI
(Juhasz 1935, 207). A stamp noted on Hadrian’s
Wall and at hinterland forts recommissioned c
AD 160; it has been noted on form 79 or
Ludowici Tg. This evidence suggests mid to lateAntonine activity, but his use of other dies on
form 27 may mean that he was active by the
middle of the second century. c AD 150–180.
Layer 2067
Severus vi 3d 33 SIIVIIRIM. Lezouxb. There is
no internal dating for this stamp, but the potter’s
output includes plain forms made at Lezoux in
the later second century, such as 31R, 79 and
80, and decorated ware which belongs
stylistically to the same period. Vessels stamped
with some of his other dies occur at northern
forts recommissioned c AD 160–c AD 160–190.
Layer 2092, KG8
Suobnedo 2b 31 SVO[BNE]DOF (Dickinson
1990, fig. 183. 41) Lezouxb. This is only the
second example noted on a dish of a stamp more
normally used on cups of form 33. Dating
evidence of Suobnedo is sparse, but he is
known to have made form 79/80, and so
Antonine activity is certain. c AD 140–170.
Layer 1055
Tituro 5A 31 [TITVR]ONIS (Dannell 1971, 315,
92) Lezouxa. One of the less-common stamps
of a potter working in the mid- to late-Antonine
period. Stamps from his other ides occur in the
Wroxeter Gutter find of the early 160s, on
Hadrian’s Wall and on some of the later 2ndcentury forms, such as 31R, 79 and 80. c AD
160–190. Layer 2012
FOSSE LANE 1996: SPECIALIST REPORTS
11
12
13
]CI (?) on form 33, burnt, Central Gaulish.
Antonine. 95/265
S [or SX [on form 31, Central Gaulish. Mid- to
late-Antonine. 80/305
(not illustrated) Frame only form 18/31.
Central Gaulish. Hadrianic/early Antonine.
?Layer 1053
Graffiti Roger Tomlin
Graffiti was recognised on four sherds (see below),
two of which are samian. This includes a deeply
scored ‘X’ on the underside of stamped vessel cat.
no. 6.
1
2
3
4
Rim sherd of a black burnished ware jar with
splayed rim, found with material of the late 3rd/
4th century. Coarsely incised after firing: two
vertical strokes close together, now incomplete,
and a third, also incomplete, overlying more or
less horizontal scoring. This could be read as
[...N]I +, the genitive case-ending of a personal
name followed by a mark of identification, but
not enough survives for certainty. Layer 2019
(Fig. 31.17)
Base sherd of a buff jar dated to the 3rd century.
Scored underneath after firing: two parallel lines
intersecting with a broad line at right angles. Not
a letter, but a mark of identification. Layer 2048
(Fig. 31.13)
Base sherd of a Dr. 33 cup stamped PATIIRATV
(Lezoux), scratched after firing within the footring: ‘+’. A mark of identification. 85/300
Rim sherd, probably of a Dr. 31 bowl (ex
catalogue), scratched after firing with four
incomplete letters just below the carination. Only
the tops of them survive, so the reading is not
certain, but possibly: [...]MPRO[...]. This could
be part of the name Sempronius, but it must be
emphasised that with these irregular capitals
being incomplete, the reading is in doubt. 85/
390
Discussion Annette Hancocks
Dating
Dating of the pottery as a whole was inhibited by its
fragmentary nature, whilst using the dating evidence
was rendered less sure by the residual character of
the majority of the assemblage. Nevertheless good
quality dating evidence was found deriving principally
from the samian and amphorae but also from Black
Burnished ware forms. The samian gave the best
evidence for 1st and 2nd-century activity. Samian
reached the site from the mid-1st century
Fig. 33 Samian stamps; scale 2:1
throughout the exporting period and into the 3rd
century. The usual peaks and troughs within the
assemblage are apparent with the bulk of the
W 31
SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011
material being of mid to late-Antonine date. The
quantity of Neronian and early Flavian material is not
great and may indicate that activity on the site did not
really begin until the 70s. For the 1990 collection
Brenda Dickinson observed that within the Southern
Gaulish decorated wares the ratio of Dr 29:Dr 37 is
about 1:3 and concluded that activity on the site may
have started as late as the mid-80s (Dickinson 2001,
144). The same ratio occurs in this assemblage.
The range of forms present is fairly standard (15/
17, 15/31, 18, 18/31, 18/31R, 27, 29, 30, 31, 31R,
32, 33, 35, 36, 37, 38, 45, 67/72, 64, 79, 79R, 80,
Curle 15 and Curle ?21). There appear to be more Dr
33 cups than Dr 27 and more Dr 31/31R than Dr 18/
31–18/31R bowls, but this was not quantified. The
Dr 64 beaker and several globular jars (including plain
examples, one with moulded and one with cut-glass
decoration) are the less common forms, but are not
unexpected in an ‘urban’ context.
A small quantity of pre-Flavian material was
identified, although the majority of the Southern
Gaulish material is Flavian or later. The latest vessel
from Southern Gaul is from Montans and probably
no later than the early 2nd century. Other early 2ndcentury products are from Les Martres-de-Veyre and
include some fine decorated bowls and plain forms
including at least one form 15/17. The bulk of the
assemblage is from Lezoux, with the greater
proportion of this material probably dating to the mid
and later Antonine period. Some of the latest vessels
to reach the site are from Eastern Gaul and include
forms 45, 32 and 36, although there are probably only
a handful of these vessels. No decorated or stamped
vessels from Eastern Gaul were identified. At least
one vessel of 3rd-century date was noted. In general
terms the trends observed here agree with those
observed for the material excavated in 1990 to the
south (Dickinson 2001), although later forms (45,
79R, 80) which were absent in that earlier assemblage
were noted here.
Trade
Some 35% of the stratified pottery was supplied by
non-local producers (Tables 5–6). Of this imported
material, BB1 was the only traded ware reaching the
site in any quantity (30% by count) with all the other
sources represented by much smaller quantities. At
the 1990 site the proportion of BB1 in the assemblage
as a whole was 37% (Evans 2001, table 3) and at
Cannards Grave 41% (Laidlaw 1997), and the figures
from both sites contrast markedly with the
representation of BB1 at Ilchester of 50–60% (Leach
1982, 142–3) and at Catsgore of 70% (Leech 1982,
159; Leach 1984, 25). This fall off in the proportion
of BB1 is thought to reflect Durotrigan ethnicity and
to be explained by Shepton Mallet’s location at a
civitas boundary (Evans 2001, 158). By mapping these
W 32
distributions, Allen and Fulford (1996) have also
suggested that the main route taken by the BB1 supply
trains from Ilchester was down river to the Severn
rather than north up the Fosse Way, and it may be that
the distribution of BB1 owes as much to the choice
of communications routes as to the preference of the
inhabitants of a civitas.
The analysis of the 1990 assemblage showed that
Severn Valley wares formed a surprisingly low
proportion of the collection (Evans 2001, 126). Not
only should the fall off in BB1 be offset by an increase
in Severn Valley ware to fit the theory that the pottery
is an indication of civitas groups, but Shepton Mallet
as a producer of Severn Valley ware should have had
high local use. However, the 1996 pottery gave a
similar figure, 9%, for the proportion of Severn Valley
ware to the 4% recorded in 1990. There was a sharp
difference in the dominance of reduced over oxidised
wares at both sites, in comparison with sites to the
north in the Severn Valley ware heartland. In the 1996
collection, greywares (including BB1 and Savernake
ware) represented 82% as opposed to oxidised wares
which represent only 5% by count. Evans has
suggested that in ceramic terms Shepton Mallet can
be best paralleled by sites north of the Mendips from
the Chew Valley to Bath. An interesting link with sites
across the Bristol Channel in Wales was also noted
(Evans 2001, 159, 160).
Study of the form/fabric occurrence in the 1996
collection and the proportion of different vessel
classes in different fabrics showed that BB1 jars and
bowls occurred in roughly equal proportions, while
tankards occurred in Shepton Mallet Severn Valley
ware along with the occasional bowl form. There was
an obvious lack of locally produced forms in the
assemblage. From this it can be inferred that the
locally produced products were traded further afield,
whilst there was a heavy reliance on the marketing of
BB1/local BBC for utilitarian forms. The
assemblage from Cannards Grave had a higher
proportion of BB1 that at the 1990 site, but this
was made up of a limited range of utilitarian forms
(Laidlaw 1997). The pottery as a whole suggested
that the local inhabitants’ access to pottery from
non-local sources was rather restricted. The
dominant supply of BB1 appeared to have been
unaffected by changes in the later centuries. This
is reflected in the continued reliance on BB1 pottery
as opposed to the New Forest and Oxfordshire
products of the 3rd and 4th centuries. Even within
the earlier phases of occupation early local products
such as Savernake ware from Wiltshire did not prevail
in any noticeable quantity. There were no New Forest
products and only a small quantity of Oxfordshire
products in the key group assemblages. As was usually
the case in the Province, imported wares were
dominated by Central Gaulish samian and a small
quantity of amphorae.
FOSSE LANE 1996: SPECIALIST REPORTS
Status/function
Overall several forms were present through the life
of the settlement including bowls (B20s), jars, jar/
cook pots, beakers and the occasional dish and tankard
form. The extent of residuality makes it difficult to
assess the chronological significance in the forms
present. With the exception of the Central Gaulish
samian, very little fineware material was recovered
from the key groups. This may be an indication that
the area excavated in 1996 was of a lower level of
status to that excavated in 1990 or may simply reflect
the type of contexts chosen. In general, looking at the
1990 and 1996 material together, a general lower
status can be suggested than in comparison to the
civitas capital at Ilchester (Leach 1982), but both
collections were of markedly higher status than at
Cannards Grave (Laidlaw 1997). Within the material
looked at in detail in 1996, few New Forest and
Oxfordshire region fine wares were found. The only
late finewares were two sherds of possibly residual
Gaulish glazed wares. This contrasts with the slight
increase of fine ware proportions seen in 1990 in the
later periods (Evans 2001, 157). The range of forms
at the Tesco site also seems smaller than at the 1990
site. Analysis of vessel class demonstrated the high
proportion of jar forms compared to other forms such
as bowls and dishes, and this was particularly the case
from Period 4 onwards. This high proportion is likely
to reflect a limited range of forms.
Generally the samian trends observed in 1996 agreed
with those presented by Brenda Dickinson for material
excavated in 1990 (Dickinson 2001) and was comparable
with assemblages from Ilchester (Leach 1982 and
1994b). All the samian from Area A was of Central
Gaulish origin (predominantly Lezoux or Les Martresde-Veyre), whilst within Area B, particularly F247, a
small amount of South Gaulish samian was observed.
The majority of samian recovered from this area was of
Central Gaulish origin. However, nearly all the decorated
samian was from Southern Gaul (La Graufesenque).
recovered from the midden F247 is significantly
heavier (31g) and may represent a primary rubbish
deposit. The Period 4 material had weights of 8g from
a floor, 13 and 8g from hearths and ovens and 12g
from a pit, all suggesting secondary deposition.
A few pieces of coarseware and samian, particularly
bowls, demonstrated evidence of prolonged use, with
lead rivets in situ or drilled holes for rivets apparent.
The same was true in the 1990 collection with the
proportion of material so treated higher than at
Ilchester (Evans 2001, 162). While the pottery in
general was in a fragmentary and abraded condition,
much of the samian was relatively unabraded although
some was heavily burnt. The condition of the pottery
in general is likely to be an indication of the degree
of disturbance on site through time with material being
re-excavated to be redeposited in the range of below
ground intrusions observed. The condition of the
samian presumably reflects its durability and its loss
and burial in the earlier centuries of occupation.
THE WORKED STONE Fiona Roe
Objects
The objects reported on here amount to several pieces
from a small millstone, six quern fragments, two
whetstones, a rubber and a small dish or mortar; none
is illustrated. Only three of these finds are from
stratified contexts. However the materials utilised, and
particularly the Mendip Old Red Sandstone, all had
long periods of use so that exact phasing for them
may not be of any great significance. Although no
stone artefacts were recorded in contexts earlier than
the late 2nd century, the likelihood is that some could
have been current earlier than this. A delay may often
have occurred between the period of use of a stone
object and its final burial as discarded rubbish, since
pieces of stone left lying on the ground surface would
often survive well when abandoned to the elements.
Taphonomy
Millstone
Variations in sherd size in different groups allows
consideration of the level of residuality and of
different kinds of artefact deposition around the site.
In Period 2, assemblages with larger mean sherd
weights were found in ditch deposits (10g and 13g)
than in material from Structure 1 (5g and 4g). These
latter are likely to have been the remnants of secondary
rubbish which had then been incorporated in floor
surfaces. The average weight of sherds from F133 of
6g indicates that the material had been redeposited in
the spread. In Period 3 the mean average sherd weight
of pottery from ditches is slightly higher at 13g than
in earlier phases but must nevertheless reflect its
arrival in the ditch after deposition elsewhere. Material
Fragments of a small millstone, no 1, made from
Upper Greensand, were reused as cobbling in the open
yard F112. Roman millstones appear to come in two
size ranges. Many were up to a metre or more in
diameter, while a smaller group varies from between
540 and 630mm in diameter (King 1988). This, and a
second millstone from the 1990 excavation, fit within
the smaller size range, with diameters of c. 580mm
for the 1996 object and 540mm for a complete lower
stone from the 1990 excavations (Roe 2001a, 235).
The pieces of Upper Greensand are now weathered,
but it can be seen that the grinding surface was
coarsely grooved. Three further unstratified greensand
fragments, no 2, may belong to the same millstone.
W 33
SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011
The greensand probably came from the Pen Pits, a
quarrying area some eleven miles (17.7km) from the
site (Pitt Rivers 1884, 6). Greensand from the Pen
Pits was used quite frequently for querns on Roman
sites in the area (Roe 2001a), despite the nearby
availability of Mendip Old Red Sandstone. At the time
of writing further examples of Roman millstones
utilising Pen Pits greensand were not known but have
since been recorded. Greensand, though unspecified,
was used for Roman millstones at Halstock, Dorset
(Lucas 1993, 96, nos 4–6), and at Figheldean, Wilts
(Mepham 1993, 36).
2000, 263). The fragment came from yard F110.
Whetstones
Both of the whetstones, nos 11 and 12, apparently
result from the casual use of broken slabs of building
stone. They are made from Old Red Sandstone, but a
dark red, micaceous variety, different from any used
for querns, and with a separate source, probably
Prior’s Hill above Wells (Green and Welch 1965, 13),
where there are old quarries some five miles (8km) from
the site. Another six Old Red Sandstone whetstones came
from the 1990 excavations (Roe 2001b).
Querns
Other stone artefacts
Old Red Sandstone from the Mendips was the most
frequently used corn-grinding material at Fosse Lane,
as might be expected with a good source for stone at
Beacon Hill only 2.25 miles (3.8 km) from the site
(Green and Welch 1965, 13; Leach 1993, 139). There
are four querns of probable Beacon Hill sandstone
from the 1996 excavations, and another six of Old
Red Sandstone from the 1990 excavations, these latter
examples dating from Period 2 (Roe 2001a). The 1996
finds of querns do not appear in the archaeological
record until the third century (Period 4). One, no 8,
which is unstratified, is a complete lower stone of the
Roman disc type, 430mm in diameter and weighing
20.8kg. This weight, for a comparatively slim disc
quern, suggests that anything larger, such as the
millstone with a diameter of 580mm, no 1, would have
needed mechanical means for turning the upper stone.
One quern fragment, no 5, is made unexpectedly of
Upper Old Red Sandstone from the Forest of Dean/
Wye Valley area, about 38 miles (61km) from the site.
The fragment was part of a quern of Roman disc type
from an undated context, the rubble from Room 2 in
Structure 7, which suggests that it may not have been
in use before the late fourth century. Imported Old
Red Sandstone could not exactly have been needed
at Fosse Lane, so its presence is something of an enigma.
The sixth quern fragment, no 6, is made from an
altogether different type of stone, a silicified Jurassic
sandstone or roach from the Harptree Beds. This is a
sandstone with casts of fossil shells, and the gaps
where they have eroded out would have provided a
good grinding surface. The nearest source for this type
of stone is at Oakhill (Green and Welch 1965, 109),
just off the Fosse Way, a short distance from Beacon
Hill, and so only 2.5 miles (4km) from the site. The
quern itself is a somewhat thick beehive type with
affinities to Iron Age varieties of quern, and this
typology is reflected in the other finds made from the
same sandstone. Beehive querns utilising the Harptree
Beds roach occurred at both the Meare and
Glastonbury Lake villages (Roe 1995, 166) and there
were also a number of similar finds from Iron Age
(Middle Cadbury) contexts at South Cadbury (Roe
W 34
Mendip Old Red Sandstone was also used for a small
rubber/hammerstone, no 9. A small dish or mortar,
no 10, only some 110–120mm in diameter, was made
from another material that was available close to the
site, shelly limestone from Doulting. More recent
quarries in the Doulting stone are sited some 2.1 miles
(3.4km) from Fosse Lane. The same stone was also
used for details on Roman buildings at Fosse Lane
(Leach 2001), and also for a child’s coffin, nos 15
and 16 (Figs 22 and 23). The small vessel probably
represents the contingent use of a spare piece of building
stone.
Building stone
It was possible to use stone extensively for building at
Fosse Lane, as good materials were available either on
or within a few miles of the site, and the local Lias in
particular was utilised in various capacities. None of the
catalogued items, as it happens, is dated earlier than the
4th century. By then (Period 5), there is some evidence,
no 13, for the use of ashlar masonry, made from the
local Jurassic limestone which was brought in from
around 2–3 miles (3.2–4.8km) south east of the site.
The remaining pieces are all from the undated late
or post-Roman period. Bath stone, which was
conveniently available up the Fosse Way, from about
16 miles (25.7km) away, was used for a column, no
14, which, with a diameter of about 170mm, was fairly
slender. Bath stone was also used for an altar from
the 1990 excavations, and seems to have been
employed generally in the area for any stonework
needing detailed shaping. At King’s Weston, for
instance, the villa was entered through a portico with
six columns of Bath stone (Branigan 1969, 19), and
the same stone was also used for architectural features
at Camerton (Wedlake, 1958, 49) and at Gatcombe
(Horwell 1977, 103).
The final item to be discussed is a stone coffin, used
for the burial of a small child, nos 15 and 16. A second
coffin, made from local Downside stone, was found
in the 1990 excavations (Leach 2001, 253). The lid
FOSSE LANE 1996: SPECIALIST REPORTS
of the child’s coffin is made from Doulting stone, which
in places outcrops adjacent to the Downside stone. In
general these Roman coffins tend to occur in areas of
suitable stone (Woodward 1993, 227), and particularly
where Jurassic limestones were available, so that finds
from Fosse Lane fit well into the known pattern.
Discussion
Sites in and around the Mendips were well provided,
within a dozen miles or so, both with good local
grinding stone and with building stone, and this
assemblage from Fosse Lane shows how the local
rocks were put to good use. However, the inhabitants
of the small town were not deterred from using
imported materials as and when the opportunity arose.
The evidence from the 1996 excavations is in some
ways misleading, since although the 1996 whetstones
and the small mortar are made of local stone, the 1990
excavations produced six well-travelled whetstones
and a mortar of Kentish Rag, and also another mortar
of Purbeck marble (Roe 2001c, 177). The Bath
freestone was probably used on many Somerset sites
for shaped or carved architectural details, so its use
at Shepton Mallet is of no surprise, but the presence
of a quern of Upper Old Red Sandstone from the
Forest of Dean is less easy to explain. One possibility
is the link with the exportation of Severn Valley Wares,
as part of the widespread trading opportunities that
were available during the Roman period.
Catalogue
1
2
3
4
5
6
Part of small millstone, diam c. 580mm, th c.
66–73.5mm, now weathered; grinding surface
was coarsely grooved, other surface pecked to
shape; three large fitting pieces. Upper Greensand,
probably from the Pen Pits. SF 386, F112
Three fragments probably of the same millstone.
Upper Greensand, probably from the Pen Pits.
SF 406 and SF 407, Area B, SF 408, Area A
Fragment upper stone rotary quern, Roman disc
type, diam c. 410mm, th at rim 48.5mm, pecked
neatly to shape, traces of grooving on grinding
surface. Mendip Old Red Sandstone
conglomerate, probably from Beacon Hill. SF
404, ditch F230
Half lower stone of rotary quern, diam c. 355mm
th at edge 36mm, in centre 77.5mm; pecked
grinding surface. Mendip Old Red Sandstone
pebbly sandstone, probably from Beacon Hill,
SF 385, F113, Structure 2
Fragment of rotary quern, Roman disc type,
probably upper stone; top surface and edge are
pecked to shape, th at edge 43mm. Upper Old
Red Sandstone from Forest of Dean. SF 293,
layer 2025, Structure 7 destruction
Half lower stone of rotary quern, diam c.
370mm, max th at edge 74mm, in centre
110mm, grinding surface worn smooth, trace
of socket for spindle. Jurassic sandstone or
‘roach’ from the Harptree Beds. SF 402, F110
7
Half upper stone of small rotary quern, diam c.
365mm, max th 60.5mm; fairly smooth grinding
surface, now weathered, trace of slot in side for
handle. Mendip Old Red Sandstone, pebbly sandstone
probably from Beacon Hill. SF 403, Area B
8
Complete lower stone large rotary quern, Roman
disc type, diam c. 430x445mm, th at rim 48mm;
grinding surface mostly worn smooth, underside
more roughly shaped, central hole min diam
43mm. Mendip Old Red Sandstone, probably
from Beacon Hill. SF 405, surface find Area B
9
Small rubber/hammerstone, utilising pebble,
78x54x25mm. Mendip Old Red Sandstone. SF
119, surface find Building 6
10 Fragment from small dish or mortar; diam c.
110–112mm, depth 45mm. Doulting stone. SF
121, surface find Building 6
11 Whetstone, utilising rectangular slab of
sandstone, wear on one flat surface, iron stained;
100x66x20mm. Dark red micaceous Mendip
Old Red Sandstone, could be from Prior’s Hill.
SF 302, layer 2030, under F228
12 Whetstone, utilising broken triangular slab, worn
smooth on four surfaces; 126x96x31mm. Dark
red micaceous Mendip Old Red Sandstone,
could be from Prior’s Hill. SF 304, 2095, rakeout
from F264
13 Three large worked fragments, burnt; two fitting,
with flat surface, the third with a flat surface
and a carved area; re-used building stone?
Oolitic limestone, coarser grained than no 14
and with some large pieces of shell; local
Jurassic. SF 292, layer 2029, in Structure 7
14 Two fragments from column, diam c. 170mm.
Oolitic limestone with a few shell fragments;
Bath stone. SF 120, Structure 6 demolition
15 Coffin lid fragment, th 98mm. Shelly limestone,
Doulting. SF342, grave F208 (Fig. 22)
16 Stone box coffin. SF 343, grave F208 (Fig. 23).
In addition a further fragment of rotary quern, SF
122, was found in Building 7 destruction. This was
not seen by the author.
THE CHARRED PLANT REMAINS Julie Jones
Introduction and methods
Extensive environmental sampling was carried out as
part of the excavation with samples recovered from
ditches, pits and other features associated with
activities in the settlement. Most of the samples were
30 litres and these were flotation sieved to a 250
micron float and 500 micron residue and allowed to
W 35
SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011
dry. Thirty-two flots which contained charred plant
macrofossils and charcoal were assessed by Vanessa
Straker (Leach and Ellis 1996). In view of the low
concentration of material in most of the samples it
was recommended that only ten were analysed further.
These were examined under a low-powered
microscope. Preservation was variable with much of
the cereal grain being in poor condition, in contrast
to the chaff and weed seeds which were generally well
preserved. The results are shown in Table 7. Plant
nomenclature and habitat information follows Stace
(1991) and cereal grain and chaff determinations are
based on Jacomet (1987).
Results
Area A
From spread F133, context 1002 (sample ES14)
Substantial quantities of barley grain were recovered,
including tail grain (smaller than 5mm) with some
examples well enough preserved to confirm the
presence of hulled barley. Only 8 wheat grains were
present with a few silicified wheat/barley awns and a
single charred oat awn. Weed seeds were also limited
to a few grasses, sedges and docks.
From ditch F134, context 1061 (sample ES29)
Cereal remains included 53 wheat and 11 barley
grains. with accompanying wheat glume bases and
spikelet forks. Glume bases of Triticum spelta
confirmed the presence of spelt wheat but most of
the glumes and spikelet forks were too abraded to
identify further. There was no accompanying chaff
with the oat grains to confirm whether these would
have been the domesticated variety or whether the
oat was present as a crop weed. Charred seeds include
typical arable weeds including orache, blackbindweed and scentless mayweed. The Brassica/
Sinapis seed may be a crop weed but could also
represent the cultivation of plants of the cabbage
family such as mustard, rape or cole. Hazelnut
fragments from this deposit may represent food debris
or could have been collected with firewood.
From burnt area F143, context 1022, within
Structure 1 (sample ES4)
This sample only contained 8 wheat grains and a
single barley grain. No chaff was present and the weed
assemblage includes a few arable weeds, with some
more typical grassland taxa including clover/medick
and ribwort plantain.
Area B
From spread of pottery and burnt material,
context 2017 (sample ES7)
This sample produced a few wheat grains with three
W 36
glume bases and a low concentration of weed seeds.
From ditch F236, context 2038 (sample ES23)
This contained four wheat grains, wheat glume bases
and spikelet forks, three oat grains and nine weed
seeds, mostly grasses.
From pit F243, context 2650 (sample ES10)
Numbers of grains were low but chaff included spelt
and wheat glume bases with two cereal sprouts. Seeds
were limited to single examples of arable weeds with
one hazel nut fragment.
From pit F243, context 2041 (sample ES9)
Limited numbers of both wheat and barley grains with
a few oats were present. Amongst the wheat chaff
some well-preserved glume bases again confirm this
as spelt. A range of weed seeds included some typical
arable weeds such as orache, black bindweed and
brome.
From ditches F245, context 2047, and F246,
context 2046 (samples ES6 and ES11)
Both ditch fills produced a few examples of wheat,
barley, a single oat grain and some wheat chaff.
Slightly higher numbers of seeds included dock,
orache, bartsia/eyebright, cleavers, stinking
chamomile as well as species typical of grassy places,
including clover/medick, vetches and grasses.
From oven F249, context 2053 (sample ES2)
This sample contained limited cereal and weed
remains.
Discussion
The range of cultivated plants and their associated
weed assemblages recovered throughout the different
phases of activity at Fosse Lane are very similar, with
both wheat and barley grains present throughout,
wheat on the whole being predominant. The single
exception is from Phase 2 where 243 barley grains
and only eight wheat grains were found amongst the
spread of burnt clay associated with F133. No cereal
chaff was found with the grains in this sample apart
from a few silicified wheat/barley awns and a single
oat awn. There were also very few weed seeds. This
is therefore likely to represent the remains of a
cleaned crop which had become burnt. None of the
grains showed signs of sprouting, although a few
detached cereal sprouts were recovered from other
samples.
On the whole the cereal grain was poorly preserved
with much of the surface layers lost through burning
and many grains were encrusted with sediment.
Although all the wheat was identified as Triticum sp,
many of the wheat grains were of the straight parallel
sided form, typical of spelt wheat and the presence of
FOSSE LANE 1996: SPECIALIST REPORTS
spelt glume bases confirms the presence of this crop.
Some grains had a more rounded dorsal surface with
steeply angled embryo characteristic of bread wheat
(Triticum aestivum), although no free-threshing wheat
chaff was recovered. Triticum spelta is the wheat most
commonly found in the Romano-British period. The
chaff recovered represents the debris from the
threshing of spelt, which does not have free-threshing
grains and would have required parching to render
the glumes brittle to free the grain.
The soils around Shepton Mallet are shallow, welldrained, brashy, calcareous clayey soils over limestone
with some deeper calcareous clayey soils. Much of
the land use today is cereals and short term grassland
with stock rearing and dairying. The environs of the
Romano-British settlement would therefore have been
suitable for local cultivation. Spelt, which is the most
commonly occurring crop recovered here, is a hardy
cereal, ideal for winter sowing, which thrives on heavy
soils. Weeds of autumn-sown grain crops (the
Secaletea) such as wheat, germinate in the autumn
and grow rapidly in the spring along with the crop
and are then harvested with the cereals. This group
of weeds includes examples recovered in many of the
samples, including stinking chamomile, cleavers and
some species of vetch. Stinking chamomile also
suggests tillage of heavier soils. Other weeds found
include black bindweed and scentless mayweed. The
oats could also have occurred as weeds, as it was not
possible to tell whether these were the domesticated
or wild variety, as no oat chaff, necessary to confirm
this, was found. The brome is also likely to have
occurred as an arable weed but may have been
collected and utilised with the crop. It has been
described as a famine food by some authors. Jones
(1981) remarks that even in this century in
Denmark, seeds of brome were collected and
ground to flour as a famine food when the rye crop
failed.
The fact that fine chaff and weed seeds have been
recovered from the Roman town does suggest that
the crops were being brought into the town partly
processed, the spelt wheat perhaps in spikelet form,
with the final cleaning and removal of grain from chaff
taking place in the settlement. However, apart from
the find of cleaned barley from F133, charred crop
and weed remains are scarce suggesting that if any
large-scale processing was taking place in the
settlement, this was in a different part of the town.
There are a number of other Romano-British sites
where plant macrofossil analysis has been carried out
in the local region. Six samples were recovered
following the excavations at Fosse Lane in 1990
(Straker 2001). These contained a similar range of
cultivated and wild plants to those recovered in 1996.
In the different phases examined in 1990 both wheat
and barley grains were present with spelt glume bases
and spikelet forks confirming the presence of spelt.
Only a limited weed assemblage was however
recovered. A burnt layer beneath a floor in a late 2nd/
3rd century building and a hearth inside a 3rd-century
building were the most productive deposits. The
hearth was composed of 70% cereal chaff, mostly spelt
spikelet forks and glume bases, with 21% grain,
largely wheat and only 9% weeds and was interpreted
as cereal chaff being used as tinder. Deposits of chaff
are frequently interpreted as fuel and have been
recorded locally at Roman sites at Catsgore (Hillman
1982) and Kenn Moor (Jones 2000) from corn drying
ovens. The layer under the floor (6044) was composed
of 69% grain, most probably spelt, with 29% largely
spelt glumes and only 2% weeds. It is not known how
the layer of burnt grain became incorporated into the
floor deposit, although this building is interpreted as
a barn, but there was no evidence that it had burnt
down.
Other Romano-British sites in the locality include
Catsgore (Hillman 1982), three sites at Ilchester
(Murphy 1982) as well as more recent excavations at
Ilchester (Stevens 1999). From the latter, a rich
assemblage of charred plant remains dominated by
grains and chaff of spelt, with only sparse barley and
free-threshing wheat was recovered. A range of arable
weeds from a number of habitats indicate the
cultivation of a range of soil types. Stevens also
suggests the cereals were stored at different times
either fully or partly processed for use by individual
households in the Roman town.
THE ANIMAL BONE Umberto Albarella and Andy
Hammon
Introduction and methods
The excavated material would seem to represent a
typical Roman urban assemblage. Cattle dominate,
followed by sheep and then pig. Horse, dog and
domestic chicken were present, but appeared in low
numbers. Fragmentation and butchery suggest that
most bones derive from food refuse. Due to its small
size it was considered inconvenient to divide the
assemblage into seven chronological groups.
Consequently, Phases 1, 2, 3 and 4, and Phases 5 and
6 have been grouped together. These will be referred
to as ‘Earlier Roman’ and ‘Later Roman/post-Roman’
respectively. Phase 7 has not been considered because
it was likely to be affected by residuality and
contamination problems. The majority of animal bone
was recovered by hand collection during the
excavation. No bulk sieving specifically for the
recovery of animal bone was undertaken. A very small
amount of animal bone came from samples taken for
flotation. The quantity of animal bone recovered from
W 37
PHASE 2
1002
2041
ES14 ES9
42
70
ContextNo
Sample No
Flotsize(ml)
CHARRED PLANT REMAINS
Cereals
Grain
Triticum sp
cf Triticum sp
Triticum sp (tail grain)
Hordeum sp
Hordeum sp (hulled)
Hordeum sp (tail grain)
Hordeum sp (hulled/tail grain)
cf. Hordeum sp
Avenasp
cf. Avena sp
Cereal indet
Total
Chaff
Triticum spelta (glume base)
Triticum sp (glume base)
Triticum sp (spikelet fork)
Triticum sp (rachis internode base)
Triticum/Hordeum sp (awns silicified)
Avena sp (awn)
Cereal sprout (coleoptile)
Total
Weeds
RANUNCULACEAE
Ranunculus spp (inside of)
Ranunculusfiammula L.
POLYGONACEAE
Fallopia convolvulus (L.)A.Love
Rumexsp
Wheat
8
Barley
180
45
12
6
Oat
13
Wheat
Wheat/Barley few
1
2047
ES11
25
PHASE 3
1022 1061
ES4
ES29
150
550
PHASE 4
2017
ES7
45
PHASE 5
2038
2053
ES23
E52
40
23
11
7
1
6
2
4
5
11
7
13
2
2
2
1
1
4
3
3
28
292
9
43
19
49
5
9
2
2
1
6
20
6
20
2
11
6
4
15
36
57
10
1
28
15
10
6
1
5
HABITAT
2650
ES10
4
2
1
4
13
3
33
116
5
18
6
13
3
7
16
2
3
2
6
3
9
13
25
5
2
1
1
Buttercup
Lesser Spearwort
Black-bindweed
Dock
2046
ES6
28
1
76
11
13
0
104
3
26
5
1
3
1
2
1
4
2
45
7
DG
M
1
19
4
2
12
4
4
1
3
1
CD
DG
SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011
W 38
TABLE 7: CHARRED PLANT MACROFOSSILS: OCCURRENCE
ContextNo
Sample No
BRASSICACEAE
Brassica/Sinapis spp
FABACEAE
Lathyrus/Vicia spp
Lathyrus/Vicia/Pisum spp
Thifohum/Medicago sPp
Vicia c.f. tetrasperma (L.)Schreber
APICACEAE
Apium graveolens L.
PLANTAGINACEAE
Plantago lanceolata L.
SCROPHULARIACEAE
Odontites/Euphrasia spp
RUBIACEAE
Galium aparine L.
ASTERACEAE
Anthemis cotula L.
Poaceaeindet
Indet
Total
OTHER REMAINS
Buds
Charcoal (fragments >2mm)
2046
ES6
2047
ES11
PHASE 3
1022 1061
ES4
ES29
Mustard/Rape/Cole
Vetch
Vetch/Pea
Clover/Medick
Smooth Tare
PHASE 5
2038
2053
ES23 ES2
HABITAT
2650
ES10
1
1
1
Wild Celery
1
Ribwort Plantain
2
CD#
1
1
2
4
3
1
1
4
1
3
1
1
Cleavers
2
Stinking
Chamomile
Knapweed
Oxeye Daisy
Scentless mayweed
2
1
2
3
1
1
2
CHSo
1
CDd
1
DG
G-rich soils
CD
1
Rush
1
4
2
1
GMRw
1
6
1
1
1
2
3
2
16
1
8
8
16
44
<20
100+
50+
G
CD
2
1
1
DG
DG#
DG
G
ws
Bartsia/Eyebright
Smooth/soft/rye
Brome
Crested Dog’s-tail
Heath-grass
Meadow-grass/
Cat’ s-tail
Grass
PHASE 4
2017
ES7
2
DG/CD
1
G
Ew sandy.peaty
G
5
1
G
3
18
1
2
26
2
1
54
13
9
31
8
3
<40
1
<200
300+
50+
<50
<10
50+
4
6
W 39
Habitats: C:
Cultivated/Arable. D: Disturbed. E: Heath/Moor. G: Grassland. H:Hedgerow. M: Marsh. R: Rivers/streams. S: Scrub. W: Woodland. d: dry soils.
n: nitrogen rich soils. o: open habitats. w: wet soils. #: cultivated plant/of economic importance
FOSSE LANE 1996: SPECIALIST REPORTS
Centaurea sp
Leucanthemum vulgare Lam
Tripleurospermum inodorum
(L.)Schultz. Bip
JUNCACEAE
Juncus sp
POACEAE
Bromus racemosus
hordaceus/secalinus
Cynosurus cristatus L.
Danthonia decumbens (L.)DC
Poa/Phleum spp
PHASE 2
1002 2041
ES14 ES9
SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011
the samples was very small and prior to the assessment
it had been incorporated into the hand-collected
material. Consequently, the hand and sieve collected
material had to be treated as a single assemblage. The
samples taken for fine sieving and flotation varied
between 20–30 litres, unless the context from which
they derived was smaller. In such cases the whole
context was sampled. The coarse residue was washed
over an approximately 2mm mesh.
The presence of residual material cannot be
estimated from the animal bone alone; a figure of
about 40% was suggested by the pottery analysis for
the later group. The later Roman/post-Roman material
has, therefore, to be treated with caution. Dogs had
gnawed 53% of the ‘countable’ post-cranial skeletal
elements. This would suggest that a fairly high
proportion of the animal bones recovered from Fosse
Lane may not result from primary deposition, but
secondary deposition caused by scavenging dogs.
Consequently spatial analyses need to be treated with
caution, although some groups, such as the
concentration of cattle mandibles in ditch F236,
context 2039, seem unaffected.
In general bone preservation (cortical integrity) was
fairly good with little exfoliation. However, some
contexts were not so well preserved, with poor cortical
integrity and extensive abrasion. There was little
variation in colour and it followed no discernable
pattern. The animal bones were heavily fragmented
and this is indicative of material deriving from
butchery and kitchen waste.
The mammal bones were recorded following a
modified version of the method described in Davis
(1992) and Albarella and Davis (1994). This system
considers a selected suite of anatomical elements as
‘countable’; it does not include every bone fragment
that is identifiable. Mandibular fragments were
considered to be ageable when there were two teeth
present with recognizable wear. Von den Driesch
(1976) defines the majority of measurements used.
All pig measurements follow the definitions of Payne
and Bull (1988).
The differentiation of sheep and goat was attempted
on the following elements: deciduous lower premolars
(dP3 and dP4); humerus; metacarpal; tibia; astragalus;
calcaneum; and metatarsal. The criteria defined by
Boessneck (1969) were used for all elements except
the teeth (Payne 1985) and the tibia (Kratochvil 1969).
The Chicken/Guinea Fowl/Pheasant (Gallus/Numida/
Phasianus) distinction was attempted on the following
elements: articular end of the scapula, shaft of the
carpometacarpus, proximal end of the femur and shaft
of the tarsometatarsus. Mandibular teeth, both in situ
and loose, were aged using wear patterns. The system
recommended by Grant (1982) was used for cattle
and pig, whereas the wear of sheep/goat teeth was
recorded according to Payne (1973; 1987).
W 40
Occurrence and frequency of species
The following mammal and bird species were
identified: Cattle (Bos taurus), Sheep (Ovis aries),
Sheep/goat (Ovis/Capra), Pig (Sus scrofa), Equid
(Equus sp.), Dog (Canis familiaris), Red Deer (Cervus
elaphus), Domestic Chicken (Gallus gallus), Chicken/
Guinea Fowl/Pheasant (Gallus/Numida/Phasianus)
and Rook/Crow (Corvus frugilegus/corone). Two bird
bones have been provisionally identified as the distal
radius of a large corvid, most likely a Raven (Corvus
corax) (context 2009) and the proximal radius of a
Woodcock (Scolopax rusticola) or a related species
(context 2050). No small mammals or fish bones were
recovered, but their absence may be due to a recovery
bias. As is well known the bones of these small
animals can easily be overlooked during excavation.
In 1990 no fish bone and only one specimen of a small
mammal – this was a Water Vole (Arvicola terrestris)
– were found (Pinter Bellows 2001).
Most of the caprine remains could not be identified
at species level and had to be recorded as sheep/goat
(Ovis/Capra). However, it may be inferred that the
majority was in fact sheep (Ovis aries), as 29
specimens belonged to this species and none could
be identified as goat (Capra hircus) – including horncores (Tables 8 and 9). However, goat must have been
present on site as nine goat specimens – compared to
56 sheep and 218 sheep/goat – were recorded from
TABLE 8: NUMBERS OF ANIMAL BONE AND TEETH NISP
(INCLUDING HAND AND SIEVE COLLECTION)
Species
Earlier
RB
Cattle (Bos taurus)
319
Sheep/Goat (Ovis/Capra)
173
Sheep (Ovis aries)
22
Pig (Sus scrofa)
36
Dog (Canis familiaris)
7
Equid (Equus sp.)
10
Red Deer (Cervus elaphus) 1
Chicken (Gallus gallus)
1
Chicken/Guinea Fowl/
Pheasant (Gallus/Numida/
Phasianus)
3
Rook/Crow (Corvus
frugilegus/corone)
Total
572
Later/
post-RB
147
55
7
25
6
11
0
1
Total
466
228
29
61
13
21
1
2
1
4
1
254
1
826
TABLE 9: ANIMAL BONE NUMBERS (NISP) AND PERCENTAGES
OF THE THREE MAJOR DOMESTICATES
Earlier
RB
Species
N
%
Cattle (Bos taurus) 319 60
Sheep (Ovis aries)* 195 36
Pig (Sus scrofa)
22
4
Total
536
Later/
post-RB
N
%
147 63
62
27
25
10
234
Total
466
257
47
770
* combines both Sheep/Goat (Ovis/Capra) and Sheep (Ovis
aries)
FOSSE LANE 1996: SPECIALIST REPORTS
the 1990 excavations (Pinter-Bellows 2001). Goat,
although at its commonest in the Romano-British period,
is never well represented in British assemblages
(Albarella and Davis 1996; Dobney et al nd).
Four different methods were employed to quantify
the relative importance of species present (Fig. 34)
They were as follows:
1
2
3
4
NISP: Number of identified specimens. A simple
count of ‘counted’ bone fragments by skeletal
element and the total number of fragments per
species.
MNI: Minimum number of individuals. The
most frequently occurring element denotes the
MNI for that species. MNI was calculated as
follows: incisors and phalanges divided by 8;
deciduous and permanent premolars by 6; M1/
2 by 4; all other elements by 2, except
metapodials. The metacarpal and the metatarsal
were calculated in the following way: metacarpal
(MC1 + MC2/2 + MC1/2 + MP2/4)/2; metatarsal
(MT1 + MT2/2 + MT1/2 + MP2/4)/21
Harcourt: Live weight ratios based on Harcourt
(1979), calculated on the meat contribution of
sheep and goats being equal to 1, pigs being
equal to 1.5 and cattle being equal to 10.
Manching: Live weight ratios based on the midpoint values taken from the Manching data set
by Dobney et al (nd). Cattle live weight was
considered to be 275 kg, sheep live weight 37.5
kg and pig live weight 85 kg. Therefore, 7.3 sheep
equals a single cow and 2.3 sheep equal 1 pig.
Fig. 34 Relative importance of the main domesticates.
Upper: Earlier RB; Lower Later/post RB
All four methods were applied to cattle, sheep and
pig, whereas horse, all other mammal and bird species
were quantified using NISP only (due to their lower
frequency). All four methods clearly demonstrate that
cattle were the most abundant species and
economically the most important animal, followed by
sheep and then pig. The frequency of the main taxa
in the Earlier Roman and Later Roman/post-Roman
periods is similar, although sheep seems slightly less
common in the later period. The pattern described
above occurs using all the methods with the exception
of MNI for Later Roman/post-Roman. This shows
cattle and sheep to be present in equal numbers,
which probably results from the small size of the
assemblage.
Unsurprisingly, the live weight estimations show
cattle to have occurred in far greater proportions for
both periods. According to this system the contribution
of beef to the meat supply would range between 84%
and 94% per cent, that of mutton between 4% and
12% and of pork between 1% and 4% (Fig. 34). The
predominance of cattle is quite common for a
Romano-British assemblage and falls within the range
of 59% cattle to 32% sheep expected on non-military
Romanised sites (King 1984). The results discussed
above are similar to those from the earlier excavation
at Fosse Lane, although in that assemblage sheep were
the most numerous species using NISP (PinterBellows 2001).
Skeletal representation and butchery
Skeletal representation has only been considered in
detail for cattle and sheep, due to the small number
of bones of all other species. In Figure 35 each cattle
and sheep skeletal element has been shown as a
percentage of the most commonly occurring element
(according to MNI). Proportions of butchered bone
and the type of butchery have been summarized in
Figure 36. For both species and periods a similar
pattern appears. This pattern generally reflects the
taphonomic processes affecting the assemblage at
Fosse Lane. Certain bones are more durable than
others, therefore dense bones, such as the calcaneum
and astragalus, survive better than less compact and
late fusing bones, such as the two ends of the femur
and the proximal humerus. For a similar reason teeth
also survive well. Brain (1981) summarizes the postdepositional destruction of skeletal elements and the
Fosse Lane cattle and sheep generally adhere to this
pattern. In addition, the recovery bias favouring the
larger bones will also lead to some elements being
under represented. The skeletal representation of cattle
and sheep at Fosse Lane does not indicate any
specialized activities, such as, for example, bone/horn
working on a commercial scale, and can entirely be
explained on the basis of recovery and taphonomic
effects.
W 41
SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011
Fig. 35 Relative abundance of skeletal elements: above cattle %MNI, below sheep and sheep/goat %MNI
Maltby (1981) has noted that the type of butchery
in cattle for the Roman period varies depending on
the nature of the site. Rural sites have prevalence for
cut bone, whereas urban sites have a prevalence for
chopped bone. At Fosse Lane (for both Earlier and
Later Roman/post-Roman) the butchery on postcranial elements was divided as 69% chopped and
31% cut. As noted above, the high level of
fragmentation is typical of kitchen and butchery waste.
The fair number of butchery marks found on the bones
also confirms this (Fig. 36).
The deposit in ditch F236 (context 2039), dated to
the Earlier Roman, contained twelve cattle mandibles
and relatively large quantities of loose teeth. Five of
the twelve mandibles had been chopped. These may
derive from primary butchery waste, therefore
suggesting that the primary dismemberment of freshly
slaughtered animals may have been carried out on site.
Three horse bones showed butchery evidence with
three cut and one chop marks present. All occur on
the lower limb bones (two first phalanges and one
metatarsal) from context 2010. Skinning probably
caused the cut marks. The first phalanx (context 2010)
displaying both cut and chop marks is more difficult
to interpret. The chop could have been made for a
variety of reasons, including a rather crude way to
separate the hide and the consumption of horse
marrow either for dogs or people. Despite the
W 42
assemblage being rather small, the skeletal
representation (Fig. 35) and butchery (Fig. 36) is
informative. Most body parts from the main
domesticates were represented; this would suggest that
animals were being butchered on site.
Fig. 36 Post-cranial proportions of butchery for
Earlier and Later periods combined (Sheep = sheep/
goat)
Age at death
The small numbers of ageable mandibles has meant
that it has only been possible to construct a
survivorship/mortality curve for the Earlier Roman
cattle (Fig. 37). The fusion data generally support the
FOSSE LANE 1996: SPECIALIST REPORTS
Fig. 37 Earlier RB: cattle survivorship curve based on
mandibles where two or more teeth are present
Age stages (O’Connor 1988):
Juvenile = M1 not in wear; Immature = M1 in wear;
M2 not in wear; Sub-adult = M2 in wear, M3 not in
wear; Adult - M3 in wear, not heavily worn; Elderly =
M3 heavily work (j or beyond sensu Grant 1982
pattern suggested by the ageable mandibles.
The Earlier Roman period saw the majority of cattle
being kept till they were adult. This would suggest
that they were primarily used for secondary
products, such as dairy or traction. Animals would
only being slaughtered once they had stopped
producing milk or they would have become too
weak to pull the plough. The fact that Roman
written sources, such as Columella and Varro,
hardly mention cow milk but at the same time
emphasize the importance of sheep and goat dairy
products (White 1970) probably suggests that cattle
were mainly used as draught animals. The lack of very
young individuals – normally associated with milk
exploitation (Payne 1973; Legge 1981) – seems to
support this suggestion.
Because of the small size of this assemblage only
very general information may be obtained from the
sheep ageing data. Sheep, for both periods, were kept
to at least one year and up to eight years (Payne’s
stages D–H) before being slaughtered. Consequently,
it is not known whether they were primarily utilized
for milk, meat, wool or a combination. Only three
very young ‘countable’ post-cranial elements were
recorded, although they were not young enough to
indicate on-site breeding.
Only very few pig mandibles were present but fusion
data suggest that two thirds of the post-cranial
‘countable’ pig bone was either unfused or fusing
(seven of eleven elements). This would suggest that
the majority of individuals were being killed while
either immature or sub-adult. This is the optimum age
for slaughter – the pig would have almost reached its
maximum weight by this age.
The number of cattle mandibles from the Earlier
Roman period has allowed a comparison with other
Roman sites (Fig. 38). The Fosse Lane survivorship
curve is comparable with other sites at Lincoln (1st–
3rd centuries), Exeter (1st–3rd centuries), and Tanner
Row, York (1st–3rd centuries) up to the sub-adult age
Fig. 38 Earlier RB: inter-site comparison of cattle
survivorship curves for Lincoln (Dobney et al 1996);
Exeter (Maltby 1979) Phases R1-R7; Tanner Row
(O’Connor 1988) Periods 3-7; for age stages see Fig.
37 caption
stage. At which point it decreases slightly, falling just
below the other curves. This may tentatively suggest
a higher reliance on beef than at the other sites.
Morphology and size
Little metric data was collected from the Fosse Lane
animal bones. This is due to the small size of the
assemblage and the high degree of fragmentation. As
a result it has not been possible to chart the changes
in size and morphology over time and a greater
emphasis has been placed on an inter-site comparison.
Maltby (1981) states that the greatest lateral length
(GLI) of the astragalus is the most commonly taken
measurement in Roman cattle assemblages.
Unfortunately, the Fosse Lane assemblage only
produced three measurements (from both periods).
Although, it has not been possible to calculate a mean
based on so few measurements, it is possible to
demonstrate that all are within the range suggested
by the other sites (Table 10). Pinter-Bellows (2001)
uses the distal breadth (Bd) of the sheep tibia to make
an inter-site comparison. Again, it has not been
possible to calculate a mean for this Fosse Lane
assemblage (only five measurable specimens from
both periods), but all are within the range suggested
by the other sites (Table 10). Thus, all that can be
said of the Fosse Lane cattle and sheep is that they
were of a similar size to animals from other Roman
sites.
Mandibular third molar’s (M ) were the only
3
element from the second Fosse Lane assemblage to
occur frequently enough to allow a reliable
comparison (Fig. 39). Length and width were
compared to specimens from sites on Ermine St in
Cambridgeshire (Albarella 1998) and from Dodder
Hill (Davis 1988). Albarella (1998) concluded that,
taking into account regional and chronological
variation, the greater size of the Ermine St specimens
may be attributed to larger imported continental cattle
making a greater contribution to the local genotype.
The Fosse Lane specimens show greater variation,
but are generally similar to the Dodder Hill cattle.
This, therefore, supports Pinter-Bellows’ (2001, 298)
W 43
SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011
TABLE 10 COMPARISON OF CATTLE AND SHEEP
MEASUREMENTS FROM SELECTED ROMANO-BRITISH SITES
Site
No
Mean Range
(1/10 mm)
(1/10 mm)
Cattle astragalus (GLI)
Carlisle, 73/74–100/1051
Corstopitum, Roman
Fosse Lane, Roman2
Exeter, 300–400
West Stow, Roman
Baylham House, 100-200
Alcester, Late Roman
Shakenoak, Late Roman
Winnall Down, Early Roman
16
9
18
18
4
10
30
454
16
Sheep and sheep/goat tibia (Bd)
31
Carlisle, 73/74–100/1051
Balksbury 1973, Roman
7
Frocester Court, 100-300
12
Exeter, 55-100
21
11
Fosse Lane, Roman2
Exeter 100-300
30
Alcester, 100-200
9
Exeter, 300-400
15
Winnall Down
8
Frocester Court, Late Roman 13
Shakenoak, Late Roman
26
Baylham House, 100-200
22
Alcester, Late Roman
59
West Stow, Roman
9
571
580
582
583
608
613
614
616
616
528-641
530-630
500-625
543-620
597-630
560-658
539-679
530-720
561-684
226
228
228
231
232
233
236
239
239
240
245
245
255
256
206-248
210-273
200-250
213-292
210-249
214-259
211-260
223-270
219-256
230-270
220-280
216-288
212-294
221-274
1
Stallibrass 1991
Pinter-Bellows 2001
All others Maltby 1981
2
Fig. 39 Inter-site comparison of Earlier RB cattle M3.
A1(M) (Albarella 1997), Dodder Hill (Davis 1988)
the second Fosse Lane assemblage seem to suggest
that this may have not necessarily been the case.
Throughout prehistoric and Roman periods it was
more normal for cattle to be raised on low ground
and sheep to be raised on higher ground (Grant 1984).
One reason for this is that sheep in low-lying areas were
prone to infestations of liver fluke (Dobney et al nd).
THE HUMAN SKELETONS Stephanie PinterBellows
Summary
hypothesis that the Fosse Lane cattle were of a
relatively small size.
Summary and conclusions
Because of the small nature of this assemblage it has
only provided general information regarding the
economy and activities of the Fosse Lane inhabitants.
In general the Fosse Lane assemblage would seem to
be very similar to that from the 1990 site (PinterBellows 2001). However, due to its small size it does
not provide much additional evidence to support the
earlier conclusions. Preservation varied, but the
majority of bones were not badly abraded or
exfoliated. This would suggest that bones were
deposited reasonably quickly. However, as suggested
by the high level of scavenger gnawing, most bones
may derive from secondary deposition. Cattle
dominate this assemblage, regardless of which method
was used to quantify species abundance (MNI, NISP,
Harcourt’s live weight ratios and the Manching live
weight ratios). This differs slightly from the relative
species abundance at the 1990 site (Pinter Bellows
2001), which has a greater number of sheep. It was
concluded that this may suggest that Fosse Lane was
not a heavily Romanised site (ibid). The findings of
W 44
A total of ten individuals were recovered. The majority
of the skeletons were found in two groupings. Stature
and the relatively low rate of pathologies show this
to be a healthy population sample.
Material and methods
A total of eight inhumations, together with a very small
amount of human bone from other contexts, was
recovered from across the site. Two of the contexts
with human bone contained articulating foetal bones
which could be shown by age and skeletal elements
not to be part of the recognised inhumations and were
counted in the total number of individuals. The
skeletons were found in two groupings and a few
isolated burials. These burials are only a small part of
the community which existed at the time, a larger
percentage were excavated in 1990. It must be kept
in mind throughout this report that the relatively small
number of skeletons means that the description of the
individual skeletons does not necessarily accurately
reflect the mortality conditions which prevailed
generally during this period.
An inventory was made of the bones and teeth from
each skeleton. Standard metric and non-metric data
FOSSE LANE 1996: SPECIALIST REPORTS
was recorded; information on sexing and ageing
characteristics was collected; and bone and dental
pathologies was noted.
The demographic characteristics of each skeleton
were established following the criteria and procedures
presented in Bass (1971), Brothwell (1981), Phenice
(1969), and Stewart (1979). Priority for gender
determination was given to innominate morphology.
Cranium morphology was also used, and, whenever
possible, supplemented by univariate measurements
of the femur and humerus head, the glenoid fossa of
the scapula, the maximum length of the talus, and
other robusticity indicators. Morphological traits of
the pelvis and cranium from skeletal series of known
sex are reported generally to be around 95% accurate
(Krogman 1962), univariate measurements range from
80 to 90% accuracy (Buikstra and Mielke 1985;
Dittrick and Suchey 1986; Steele 1976). Sexing was
only attempted for adult skeletons (a term used here to
indicate those above the age of approximately 20 years).
Sub-adult age was determined through dental
development (Logan and Kronfeld as presented in
Downer 1975) diaphysis lengths (Scheuer et al. 1980;
Workshop of European Anthropologists 1980) and
epiphyseal union (Krogman 1962; Brothwell 1981).
Adult age was evaluated using the recommendations
of the Workshop of European Anthropologists (1980)
for pelvis and cranium, Suchey et al (unpub) also for
pubic symphysis metamorphosis, auricular surface
metamorphosis by Lovejoy et al. (1985) and dental
attrition (Miles 1962; Brothwell 1981). A range of
possible ages was first established, using all indicators
applicable and then a final best estimate of age was
determined by the smallest range of agreement among
the indicators. The adults here have been separated
into three groups, Young adults (20–29.9), Middleaged adults (30–49.9), and Old adults (50+). As
children’s maturation is considered to be a more
regular process than the degeneration of the adults’,
the sub-adults have been divided into more age
categories: Foetal – Birth, Birth – 0.9 years, 1–4.9,
5–9.9, 10–14.9, 15–19.9. One other category is also
used, Adults age unknown.
Measurements were taken following descriptions
in Bass (1971) and Brothwell (1981). Stature was
calculated following Trotter (1970). Cranial nonmetric variants were recorded as described by Buikstra
(1976) and the post-cranial non-metric variants
following Finnegan (1978). Pathological conditions
were evaluated through gross anatomical observation
and radiographic examination. Criteria for probable
diagnosis stemmed from Steinbock (1976), Ortner and
Putschar (1981) and Rogers et al. (1987). Coding for
dental pathologies followed Hillson (1979).
Condition and taphonomy
The preservation of the majority of the skeletal
TABLE 11: CONDITION AND DEGREE OF COMPLETENESS OF
SKELETONS
<20%
20-40%
40-60%
60-80%
80%+
Total
Good
2
2
2
6
Condition
Fair
1
1
2
4
Total
Poor
0
3
0
1
2
4
10
material was good (Table 11). Preservation was
scored as good, fair or poor on the basis of a visual
inspection of the remains. Sixty percent were in good
condition, 40% fair and none in poor condition.
The degree of completeness varied. Forty percent
were over 80% complete, but 30% has less than a
fifth of the skeleton present. The less complete
skeletons are more often found in the shallow graves
and those disturbed by ploughing. Both preservation
and completeness are very similar to that observed
for the skeletons from the 1990 excavations.
Sex and age
Of the six adult inhumations, five had characteristics
allowing the sex to be determined. Metric standards
developed during analysis of the 1990 skeletal
material (Pinter-Bellows 2001) were used for
univariate measurements of the femur and humerus
head, the glenoid fossa of the scapula, and the
maximum length of the talus. Two skeletons were
diagnosed as female and three as male (Table 12).
This fairly even gender ratio was also seen in the 1990
excavation. It is generally assumed that cemeteries
with an even sex distribution are likely to be those
where a representative selection of the whole adult
population was buried, probably in family groups.
This certainly appears to be the case archaeologically
and osteologically at Fosse Lane.
TABLE 12: DEMOGRAPHY
Age
Foetal -birth
Birth -.9
1-4.9
5-9.9
10-14.9
15-19.9
Young adults
(20-29.9)
Middle-aged adults
(30-49.9)
Old adults
(50+)
Adults age
unknown
?
3
1
Total
5
Sex
M
3
F
Total
3
1
1
1
1
4
1
1
3
2
10
W 45
SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011
In turning to the age profile of this small group
(Table 12), 40% of the individuals are sub-adults
(those individuals approximately under the age of 20
years). This figure is around that which would be
expected in a representative population sample. Three
of the four sub-adults recovered are late foetal or
neonate. It was not possible from the skeletal evidence
to determine whether they were stillborn or died in
the immediate post-natal period. The other sub-adult
died within the first year of life. In a larger grouping
of skeletons where statistical chance figures less
heavily into the age of death, the skeletons of several
older children would have been expected. The adult
age distribution shows that the greatest percentage of
aged adults were in the middle-aged category (four
out of the five adults which could be aged). This is a
quite common distribution.
Stature
The four adults whose stature could be calculated fit
comfortably within the distribution from the 1990
excavation (Table 13). Developmental stress (which
can affect stature) does not seem to have been a
problem for Fosse Lane with the mean statures for
both males and females equal to the tallest means of
the other population samples (Pinter-Bellows 2001).
This similarity of stature shows that the sub-adults
experienced no more overall stress during
development than the sub-adults from other sites
examined.
Skeletal and dental malformations and anomalies
One minor skeletal anomaly was noted. The dens
epiphysis of the 2nd cervical vertebra never fused to
the body of the dens in the middle-aged male (HB 8)
leading to a pseudarthosis (a supernumerary
articulation). This appears to have caused little trouble
for the individual, the articulation on the 1st cervical
vertebra is just slightly larger and on two planes, and
while slightly rough does not show much osteophyte
development.
Pathology
Dental disease
For the six adults, four maxillae and four mandibles
survive, with all tooth positions available for study.
Only one of these four individuals, a middle-aged male
(HB 8), shows evidence of ante-mortem tooth loss.
Two individuals have caries, a middle-aged female
(HB 1) with four and a middle-aged male (HB 8) with
five. Caries are formed when sugars in the presence
of harmless bacteria ferment and the resultant plaque
has a pH low enough to demineralise enamel, cement,
and dentine. In the male one of these caries has
resulted in a point abscess, a local circumscribed area
of infection (osteomyelitis).
Two individuals, a middle-aged female (HB 1) and
a middle-aged male (HB 7) have dental calculus.
Dental calculus may be considered as a mineralized
dental plaque. It takes the form of a concretion on the
teeth consisting mainly of calcium salts and, in life,
organic material in which flourishes numerous
bacteria; it is associated with poor oral hygiene.
Periodontal disease is perhaps the best skeletal
indicator of general oral hygiene (Manchester 1984,
34). It involves inflammatory pitting of the alveolar
margins and the progressive alveolar resorption
resulting in exposure of tooth root. Two individuals
show evidence of this, a young adult female (HB 2)
and a middle-aged male (HB 7).
Enamel hypoplasia is an area with a deficiency of
enamel of a developmental origin in a tooth. The area
affected is usually a band or a line of pitting. It occurs
during acute severe stress and has a non-specific
aetiology. It is associated with general systemic
disorder and nutritional deficiency. There are also
some forms of enamel hypoplasia which are hereditary
or occur through trauma to the tooth; in order to
identify and discount these types, in this study enamel
hypoplasia was only counted if at least two teeth from
different classes were affected. It was noted on
permanent teeth of both of the females (HB 1 and
HB 2).
TABLE 13: STATURE, MEANS AND RANGES
Sex
N
Measurements
Tesco (1996)
Female
Male
2
2
157 cm (5' 1") & 159 cm (5' 2")
174 cm (5' 7") & 180 cm (5' 8")
Sex
N
Mean
Range
162 cm (5' 4")
171 cm (5' 7")
157 cm (5' 2") - 168 cm (5' 6")
162 cm (5' 4") - 178 cm (5' 9")
Showerings (1990)
Female
10
Male
12
*n is the number of individuals for which the elements were observable
W 46
FOSSE LANE 1996: SPECIALIST REPORTS
Metabolic and nutritional disease
One individual, an infant whose tooth eruption/
calcification was of a nine month old ± three months
(HB 5), has porotic hyperostosis. It is presented as a
response to chronic anaemia (Moseley 1974). This
anaemia can be caused by a number of factors, such
as physiological susceptibility, poor nutrition or
parasitic infestation. In this individual, the associated
anaemia would probably be iron deficiency in
childhood probably caused by malnutrition and
infections; an iron inadequacy can also be exacerbated
by high lead intake. The anaemic state stimulates a
proliferative reaction of the marrow in an attempt to
make good its deficiency. The marrow space enlarges
at the expense of the outer layer of bone. This becomes
thin and porous and the marrow cells extend through
it. The subsequent deposition of new bone upon these
cells has given rise to the classic ‘hair-on-end’
appearance of the outer table in this individual of the
frontal, parietals and occipital. The outer surface of
the long bones is also raised and very porotic. In this
individual the chronic anaemia has lead to a ‘failure
to thrive’ (obviously followed by death) as while the
dental age of this individual is estimated at
approximately nine months, measurements of the long
bones gives a developmental age of a 40 week ± two
week old foetus.
An aetiologically related condition, cribra orbitalia,
has been observed of the trabecular type in a middleaged male (HB 7). Cribra orbitalia is an increase of
the diploic bone in the orbital roofs resulting in an
increase in the thickness of the orbital plate and sievelike lesions or pits appearing in the thin cortical bone
layer of the orbital roofs. It is usually considered to
be due to anaemia in young childhood, and not to
develop in response to anaemia in adulthood (StuartMacadam 1985, 398); so this individual survived the
early anaemia, malnutrition and infections which
killed the infant.
Infection and inflammation
The majority of infections affect the soft tissues of
the body, and many, from influenza and measles to
meningitis and pneumonia, run their course too rapidly
for the infection process to spread to the bones. Bone
lesions are, therefore, going to represent the chronic
infections, which are more likely to involve a bacteria
than a virus, viral infections being resolved more
rapidly. The Fosse Lane individuals show only
evidence of non-specific infections, inflammatory in
nature, for which the pathogenic agent is unknown.
Lesions which are superficial and appear to involve
only the fibrous covering of the bone (the periosteum)
are termed periostitis. Periostitis is recognised as a
deposition of irregular new bone upon the outer
surface of a bone. It is usually thin and localized in
area, but can be thicker and cover a bone more
extensively.
Three individuals exhibit periostitis. One young
adult female (HB 2) has unhealed periostitis in a line
21mm long by 9mm wide perpendicular to the spine
on both sides of the body of the left scapula. The
sternal end of ten ribs, both left and right have porous
new bone on the external surface. Two middle-aged
males have healed periostitis on the lower leg; one
(HB 7) on the left tibia and fibula and one (HB 8) on
the left fibula. Periostitis, in the absence of general
pathology, has been noted in association with repeated
and minor trauma to the lower legs (Manchester
1984), chronic venous insufficiency (Resnick and
Niwayama 1988) and tropical ulcers with soft tissue
infection (Molleson 1993).
Trauma
Trauma may affect soft tissue, bone, or both. For
trauma to the bone – fractures – there is direct
evidence (as discussed below). However, the majority
of injuries are sustained by the soft tissues. Evidence
for these injuries is indirect and depends on the severity
of the damage to the soft tissues adjacent to the bone.
There are four individuals who have injuries to their
entheses or syndesmoses. The osseous sites of
tendons, which attach the muscles to the bone, and
ligaments, which attach near particular joints and limit
movement, entheses and syndesmoses respectively,
can have alterations occur in the fibres. Disruption of
the fibre bundles that are anchored to the underlying
bone can cause hyperostotic (bone growth) or
osteolytic changes (bone loss) to that area of the bone.
They are often degenerative in nature and are
‘common in older individuals’ (Resnick and
Niwayama 1981, 1297). If the individual in which
the lesion occurs is young and robust and disease is
not suggested as the causative agent, then direct and
work-related trauma or loading stresses may be explored.
The middle-aged female (HB 1) has 1mm high
enthesophytes (extra bone) on the proximal hand
phalanges where the flexor digitorum superficialis
attach. This is seen on the proximal and medial
phalanges of the right but not the left hand on a
middle-aged male (HB 1). The young adult female
(HB 2) has enthesophytes on the left and right humeri
on the distal, anterior surface where the brachialis
attaches and on the proximal posterior surface of the
right fibula where the tibialis posterior attaches. She
also has a small roughened circumscribed
enthesolithic lesion (area of bone loss) on the left
femur on the distal posterior where the medial head
of the gastrocnemius attaches. The adult of unknown
sex and age (HB 4) has an enthesophyte on the right
ulna on the proximal end where the anconeus attaches;
a middle-aged male (HB 8) has an enthesophyte for
the small muscle of the left ulna.
W 47
SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011
Another pathology thought to sometimes originate
from stress to soft tissue, in this case the intervertebral
discs, is Schmorl’s node. While its aetiology is not
completely understood, it is believed that if the disc
located between the vertebrae is subject to too much
strain it may rupture. The bubble of escaped material
then presses against the body of the adjacent vertebra,
which gradually yields to the pressure, allowing a
small cavity to be formed in its body. Schmorl’s nodes
provide evidence of torsional and compressional
injuries that have occurred as a result of the sudden
application of loading stresses if they are not
degenerative or related to various diseases (Resnick
and Niwayama 1981, 1404). One middle-aged male
(HB 3) exhibits these nodes on five mid to lower
thoracic vertebrae.
The one bone fracture found was probably
accidental (using criteria from Manchester 1983, 58).
A middle-aged male (HB 3) has a healed, simple,
closed oblique fracture to the tip of the medial
malleolus of the left tibia with less than 5mm of
proximal-anterior displacement; and a simple, closed
oblique fracture of the distal articular end of the left
fibula with slight displacement.
Neoplastic disease
Neoplasma or new growth may be classified as
malignant or benign. No cases of malignant neoplasms
have been found in this sample. There is one case of
a benign neoplasm, an osteoma in bone cells in
circumscribed areas, normally in the periosteum,
which grow more than the surrounding tissue. The
cells eventually mature and the osteoma is not
progressive. They are small hemispherical hard
projections which are usually about the size of a pea.
They are usually symptomless and are not considered
to be of clinical significance. A middle-aged male (HB
8) has one on the proximal right thumb phalange, just
distal of midway on the medial side of the shaft.
Conclusions
The sample gives a look at what appears to be two
family groups and a few others from part of a small
community. These individuals were relatively healthy.
The adult skeletons are of medium height with mean
statures for the males close to the means for other
Roman populations in Britain and the females slightly
taller than the means of many. There are not a high
number of osseous pathologies.
REFERENCES
Albarella, U, 1998, ‘The animal bones’, in P Ellis et
al., Excavations Alongside Roman Ermine Street,
Cambridgeshire, 1996, BAR 276, 99–104
W 48
Albarella, U., and Davis, S., 1994, The Saxon and
Medieval Bones Excavated 1985–1989 from West
Cotton, Northamptonshire, English Heritage,
Ancient Monuments Lab Report 17/97
_____ , 1996, ‘Mammals and birds from Launceston
Castle, Cornwall: decline in status and the rise in
agriculture’, Circaea, 12 (1), 1–156
Allason-Jones, L., 1989, Earrings in Roman Britain,
BAR 201, Oxford.
Allason-Jones, L., and Miket, R., 1984, The
Catalogue of Small Finds from South Shields
Roman Fort, Soc. Antiq. Newcastle Upon Tyne
Monograph Series No. 2.
Allen, J.R.L., and Fulford, M.G., 1996, ‘The
distribution of south-east Dorset Black Burnished
category I pottery in south west Britain, Britannia,
27, 223–81.
Bagwell, T., and Webster, C., 2005. ‘Somerset
Archaeology, 2004’, SANH 148, 114–15.
Barrett, J.C., Freeman, P.W.M., and Woodward, A.,
2000, Cadbury Castle, Somerset: The Later
Prehistoric and Early Historic Archaeology,
English Heritage Archaeol Rep 20.
Bass, W., 1971, Human Osteology: A Laboratory and
Field Manual of the Human Skeleton, Columbia.
Bevan, L., 1998, ‘Bracelets; pins’, in P. Leach, Great
Witcombe Roman Villa, Gloucestershire. A Report
on Excavations by Ernest Greenfield 1960–1973,
BAR 266, 86–90.
_____ , 2001a, ‘The flint’, in Leach 2001, 100–7.
_____ , 2001b, ‘Fired clay, brick and tile’, in Leach
2001, 225.
Biddle, M., 1967, ‘Two Flavian burials from
Winchester’, Antiq Journ, 47, 224–50.
Birbeck, V., 1997, Excavations on Iron Age and
Romano-British Settlements at Cannards Grave,
Shepton Mallet, Wessex Archaeol Rep. 38985.
_____ , 2002. ‘Excavations on Iron Age and RomanoBritish settlements at Cannards Grave, Shepton
Mallet’, SANH 144, 41–116.
Boessneck, J., 1969, ‘Osteological differences
between sheep (Ovis aries Linné) and goat (Capra
hircus Linné), in D. Brothwell and E. Higgs (eds),
Science in Archaeology, London, 331–58.
Booth, P., 1982, ‘A Romano-British burial from
Mancetter’, Trans Birmingham Warks Archaeol
Soc, 92, 134–6.
Brain, C.K., 1981, The Hunters or the Hunted?
Introduction to Taphonomy, London
Branigan, K., 1969, The Romans in the Bristol Area,
Bristol.
_____ , 1977, Gatcombe: The Excavation and Study
of a Romano-British Villa Estate, 1967–76, BAR
44, Oxford.
Broomhead, R.A., 1999. ‘Ilchester Great Yard
archaeological excavations 1995’, SANH 142,
139–91.
Brothwell, D., 1981, Digging Up Bones, 2nd edn,
FOSSE LANE 1996: SPECIALIST REPORTS
London.
Buikstra, J., 1976, Hopewell in the Lower Illinois
Valley – a Regional Study of Human Biological
Variability and Prehistoric Mortuary Behavior,
Evanston, Illinois.
Buikstra, J., and Mielke, J., 1985, ‘Demography, diet,
and health’, in The Analysis of Prehistoric Diets
(R Gilbert Jr and J Mielke eds), Orlando, Florida,
359–422.
Burnham, B.C., and Wacher, J.S., 1990, The Small
Towns of Roman Britain, Berkeley.
Calza, G., and Becatti, G., 1977, Ostia, Rome.
Coles, J., and Minnitt, S., 1995, ‘Industrious and
Fairly Civilised’: The Glastonbury Lake Village,
Taunton.
Cool, H.E.M., 1990, ‘Roman metal hair pins from
southern Britain, Archaeol Journ, 147, 148–82.
Cool, H.E.M., and Price, J., 1985, ‘Glass’, in H. Hurst,
Kingsholm, Gloucester, 41–54.
_____ , 1991, ‘The Roman vessel and window glass’,
in T. Padley, The Metalwork, Glass and Stone
Objects from Castle St, Carlisle: Excavations
1981–2, 165–76.
_____ , 1993, ‘Roman glass’, in P. Woodward et al.,
Excavations at the Old Methodist Chapel and
Greyhound Yard, Dorchester, 1981–1984,
Dorchester, 150–67.
_____ , 1995, Roman Vessel Glass from Excavations
in Colchester, 1971–85, Colchester Archaeol Rep,
8.
Corney, M., 1997, ‘The origins and development of
the ‘small town’ of Cunetio, Mildenhall,
Wiltshire’, Britannia, 28, 337–50.
Costen, M., 1992, The Origins of Somerset,
Manchester.
Crabtree, P.J., 1989, West Stow Early Anglo-Saxon
Animal Husbandry, East Anglian Archaeology, 47.
Crummy, N., 1983, The Roman Small Finds from
Excavations in Colchester 1971–9, Colchester.
Dannell, G.B., 1971, ‘The samian pottery’, in B
Cunliffe, Excavations at Fishbourne 1961–1969,
Leeds, 260–318.
Dannell, G.B., and Wild, J.P., 1987, Longthorpe II.
The Military Works-Depot: An Episode in
Landscape History, Britannia Monographs 8.
Davey, J.E., 2005. The Roman to Medieval Transition
in the Region of South Cadbury Castle, Somerset,
BAR Brit Ser 399, Oxford.
Davis, S., 1988, Animal Bones from Dodder Hill, a
Roman Fort Near Droitwich (Hereford and
Worcester) Excavated in 1977, English Heritage
Ancient Monuments Lab Rep 140/88.
_____ , 1992, A rapid method for recording
information about animal bones from
archaeological sites, English Heritage Ancient
Monuments Lab Rep19/92.
Déchelette, J., 1904, Les Vases Céramiques Ornés
de la Gaule Romaine, Paris.
Dickinson, B., 1990, ‘The samian ware’, in McCarthy
1990, 213–36.
_____ , 2001, ‘Samian ware’, in Leach with Evans
2001, 144–9.
Dittrick, J., and Suchey, J., 1986, ‘Sex determination
of prehistoric central California skeletal remains
using discriminant analysis of the femur and
humerus’, American Journ of Physical
Anthropology, 70, 3–9.
Dobney, K., Jaques, D., and Irving, B., undated, Of
Butchers and Breeds. Report on Vertebrate
Remains from Various Sites in the City of Lincoln,
Lincoln Archaeological Studies 5, Lincoln.
Down, A., and Rule, M., 1971, Chichester
Excavations 1, Chichester.
Downer, G.C., 1975, Dental Morphology, Bristol.
Driesch, A. Von den, 1995, A Guide to the
Measurement of Animal Bone from Archaeological
Sites, Peabody Museum Bulletin 1, Cambridge
Mass., Harvard University.
Ellis, P., 1984. Catsgore 1979, Further Excavation
of the Romano-British Village, Bristol.
Ellis, P. and Leach, P., 2000 The Roman small town
at Shepton Mallet, Somerset. The Tesco
Excavation 1996, Birmingham University Field
Archaeology Unit No. 449.
Esmonde Cleary, S., 2001. ‘Roman coins’ in Leach
with Evans 2001, 211–24.
Evans, C.J., 2001. ‘The Roman pottery’, in Leach with
Evans 2001, 107–69.
Farwell, D.E., and Molleson, T.I., 1993, Poundbury
Volume 2, The Cemeteries, Dorchester.
Ferris, I.M., and Cooper, J., 1996, ‘Copper alloy
objects’, in I. Ferris and A. Esmonde-Cleary,
Excavations at New Cemetery, Rocester,
Staffordshire, 1985–1987, Trans Staffs Arch Hist
Soc, 35.
Finnegan, M., 1978, Non-metric variation of the
infracranial skeleton, Journ of Anatomy, 125, 23–
37.
Fowler, P., and Rahtz, P., 1972. ‘Somerset AD 400–
700’, in P. Fowler (ed), Archaeology and the
Landscape, Bradford on Avon, 187–221.
Fremersdorf, F., and Polonyi-Fremersdorf, G., 1984,
Die farblosen Gläser der Frühzeit in Köln 2 und
3 Jahrhundert, Denkmäler des römischen Köln 9.
Gillam J.P., 1976, ‘Coarse fumed ware in northern
Britain’, Glasgow Archaeol Jnl, 4, 57–80.
Grant, A., 1982, ‘The use of tooth wear as a guide to
the age of domestic ungulates’, in B. Wilson et al.
(eds), Ageing and Sexing Animal Bones from
Archaeological Sites, BAR 109, Oxford, 91–108.
_____ , 1984, ‘Animal husbandry in Wessex and the
Thames valley’, in B. Cunliffe and D. Miles (eds),
Aspects of the Iron Age in Central Southern
Britain, Oxford, 102–19.
Graham, A., and Newman, C., 1993, ‘Recent
excavations of Iron Age and Romano-British
W 49
SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011
enclosures in the Avon Valley, Wiltshire’, Wiltshire
Archaeol Magazine, 86, 8–57.
Green, M.A., and Welch, F.B.A., 1965, Geology of
the Country around Wells and Cheddar,
Geological Survey of Great Britain, Memoir for
Sheet 280, London.
GSA 1990a. Report on Geophysical Survey: Mendip
Business Park, Shepton Mallet, Report 90/51,
Geophysical Surveys Bradford.
GSA 1990b. Report on Geophysical Survey: Fosse
Way, Shepton Mallet, Report 90/52, Geophysical
Surveys Bradford, 1990b.
Guido, M., 1978, The Glass Beads of the Prehistoric
and Roman Periods in Britain and Ireland, Res
Rep Soc Antiq 35, London.
Harcourt, R.A., 1979, ‘The animal bones’, in G
Wainwright (ed), Gussage All Saints: An Iron Age
Settlement in Dorset, London, 150–60.
Harden, D.B., 1961, ‘Domestic window glass: Roman, Saxon and Medieval’, in E. Jope, Studies in
Building History, London, 49–52.
_____ , 1983, ‘The glass hoard’, in S. Johnson, Burgh
Castle. Excavations by Charles Green 1958–61,
East Anglia Archaeology Report 20, 78–89.
Hartley, K.F., 1991, ‘The mortaria’, in Holbrook and
Bidwell 1991, 189–215
_____ , 2001, ‘Shepton Mallet mortaria’, in Leach
2001, 130–2.
Hillman, G., 1982, ‘Evidence for spelt malting at
Catsgore’, in Leech 1982, 137–41.
Hillson, S.W., 1979, ‘Diet and dental disease’, World
Archaeol, 11, 147–62.
Hoffmann, B., 1996, Römisches Glas in BadenWürttemberg von 70 bis 260, unpub Phd thesis,
Freiburg University.
_____ , forthcoming, The Roman Glass from
Vindolanda, Vindolanda Research Reports.
Holbrook, N., and Bidwell, P.T., 1991, Roman finds
from Exeter. Exeter Archaeol Report 4.
Hooke, D., 1989. ‘Early medieval estate and
settlement patterns’, in M Aston et al., The Rural
Settlements of Medieval England, London.
Hopstätter, H., 1942, ‘Ein römisches Brandgrab bei
Kisselbach (Hunsrück)’, Germania, 26, 211–13.
Horwell, D., 1977, Stone objects, in Branigan 1977,
99–102.
Jacomet, S., 1987, Prahistorische Getreidefunde.
Botanisches Institut der Universitat Abteilung
Pflanzensystemematik und Geobotanik, Basel.
Jones, J., 2000. ‘Plant macrofossils’, in S. Rippon,
‘The Romano-British exploitation of coastal
wetlands: survey and excavation of the North
Somerset Levels, 1993–7’, Britannia, 31, 122–38.
Jones, M., 1981, ‘The development of crop
husbandry’, in M. Jones and G. Dimbleby, The
Environment of Man: the Iron Age to the AngloSaxon Period. BAR British Series 87.
Juhasz, G., 1935, Die Sigillaten von Brigetio,
W 50
Dissertationes Pannonicae, Ser 2, no 3,
Budapest.
Kilbride-Jones, H.E., 1938, ‘Glass-armlets in Britain’,
Proc Soc Antiq Scotland, 72, 366–72.
King, A.C., 1984, ‘Animal bones and the dietary
identity of military and civilian groups in Roman
Britain, Germany and Gaul’, in T. Blagg and A.
King (eds), Military and Civilian in Roman
Britain: Cultural Relationships in a Frontier
Province, Oxford, BAR 136, 187–218.
King, D., 1998, Mechanised corn milling in Roman
Britain, Quern Study Group.
Kratochvil, Z., 1969, ‘Species criteria on the distal
section of the tibia in Ovis ammon F. Aries L. and
Capra aegrus F. Hircus L.’, Acta Veterinaria, 38,
483–90.
Krogman, W., 1962, The Human Skeleton in Forensic
Medicine, Springfield, Ill.
Laidlaw, M., 1997, ‘Romano-British pottery’, in
Birbeck 1997, 24–32.
Lawson, A.J., 1975, ‘Shale and jet objects from
Silchester’, Archaeologia, 105, 241–75.
Legge, A.J., 1981, ‘The agricultural economy’, in R.
Mercer, Grimes Graves, Norfolk: Excavations
1971–72, DoE Res Rep 11, London, 79–103.
Leach, P.J., 1982. Ilchester Volume 1: Excavations
1974–5, Bristol.
_____ , 1984, ‘The pottery, in Ellis 1984, 23–8.
_____ , 1990. An Archaeological Assessment of the
Mendip Business Park, Fosse Lane , Shepton
Mallet 1990, BUFAU.
_____ , 1991. An Archaeological Evaluation at
Bullimore Farm, Shepton Mallet, Somerset 1991,
BUFAU.
_____ , 1992. An Archaeological Evaluation of
Development Land at Fosse Lane, Shepton Mallet,
Somerset 1992, BUFAU.
_____ , 1994a. Mendip Business Park, Shepton
Mallet: Site A 4/94, Archaeological excavations
in advance of development, BUFAU Report 298.
_____ , 1994b. Ilchester Volume 2: Archaeology,
Excavations and Fieldwork to 1984, Sheffield.
_____ , 2001. Roman Somerset, Dorchester.
_____ , 2009. ‘Prehistoric ritual landscapes and other
remains at Field Farm, Shepton Mallet’, SANH
152, 11–68.
Leach, P.J., and Ellis, P., 1996. Fosse Lane, Shepton
Mallet, the Tesco excavations, BUFAU interim
report.
Leach, P.J, with Evans, C.J., 2001. Excavation of a
Romano-British Roadside Settlement in Somerset:
Fosse Lane, Shepton Mallet, 1990, Britannia
Monograph 18, London.
Leech, R.H., 1982. Excavations at Catsgore 1970–
1973, Bristol.
Lovejoy, C.O., Meindl, R., Mensforth, R., and Barton,
T., 1985, ‘Chronological metamorphosis of the
auricular surface of the ilium: a new method for
FOSSE LANE 1996: SPECIALIST REPORTS
the determination of adult skeletal age at death’,
American Journl of Physical Anthropology, 68,
15–28.
Lucas, R.N., 1993, The Romano-British villa at
Halstock, Dorset: Excavations 1967–1985, Dorset
Nat Hist and Archaeol Soc Mono Series 13.
Mackreth, D., 2001, ‘Brooches’, in Leach 2001, 179–
200.
Maltby, M., 1979, The Animal Bones from Exeter
1971–1975, Sheffield: Exeter Archaeological
Reports 2.
_____ , 1981, Iron Age, Romano-British and AngloSaxon animal husbandry: a review of the faunal
evidence, in M Jones and G Dimbleby (eds) The
Environment of Man: The Iron Age to the AngloSaxon Period, BAR 87, Oxford, 254–61.
Manchester, K., 1983, The Archaeology of Disease,
Bradford.
_____ , 1984, Eccles skeletal report, unpub.
Manning, W.H., 1985, Catalogue of the RomanoBritish Iron Tools, Fittings and Weapons in the
British Museum, London.
McCarthy, M., 1990, A Roman, Anglian and Medieval
site at Blackfriars Street, Carlisle, Cumb West
Archaeol Soc Res Series 4.
Mepham, L., 1993, ‘The stone’, in Graham and
Newman 1993, 36.
Miles, A.E.W., 1962, ‘Assessment of the ages of a
population of Anglo-Saxons from their dentitions’,
Procs Royal Society of Medicine, 55, 881–6.
Molleson, T.I., 1993, ‘The human remains’, in Farwell
and Molleson 1993, 142–213.
Morgan, G., 2001, ‘Mortar and plaster analysis’, in
Leach 2001, 226–30.
Moseley, J.E., 1974, ‘Skeletal changes in the
anemias’, Seminars in Roentgenology, 9(3), 169–
84.
Murphy, P., 1982, ‘Plant remains from Roman
deposits at Ilchester’, in Leach 1982, 286–90.
Neale, F., 1976. ‘Saxon and medieval landscapes’, in
R Atthill (ed) Mendip, A New Study, Newton
Abbot, 75–101.
O’Connor, T., 1988, ‘Bones from the General
Accident Site, Tanner Row’, Archaeology of York
fascicule 15/2, The Animal Bones, CBA Res Rep,
London.
Ortner, D., and Putschar, W., 1981, Identification of
Pathological Conditions in Human Skeletal
Remains, Smithsonian Contributions to
Anthropology 28, Washington, DC.
Oswald, F., 1936–7, Index of Figure-Types on TerraSigillata, Liverpool.
Payne, S., 1973, ‘Kill-off patterns in sheep and goats’,
Anatolian Studies: Journal of the British Institute
of Archaeology at Ankara, 23, 281–303.
_____ , 1985, ‘Morphological distinctions between
the mandibular teeth of young sheep, Ovis, and
goats, Capra’, Journal of Archaeological Science,
12, 139–47.
_____ , 1987, ‘Reference codes for the wear states in
the mandibular cheek teeth of sheep and goats’,
Journal of Archaeological Science, 14, 609–14.
Payne, S., and Bull, G., 1988, ‘Components of
variation in measurements of pig bones and teeth,
and the use of measurements to distinguish wild
from domestic pig remains’, Archaeozoologia, 2,
27–65.
Pengelly, H., 1982, ‘The samian;, in J. Draper and C.
Chaplin, Dorchester Excavations Volume 1:
Excavations at Wadham House 1968; Dorchester
Prison 1970, 1975, and 1978; and Glyde Path
Road 1966, Dorset Natur Hist Archaeol Soc Mono
2, 78–9.
Phenice, T., 1969, ‘A newly developed visual method
of sexing the os pubis’, American Journ of
Physical Anthropology, 30, 297–301.
Philpott, R., 1991, Burial Practices in Roman Britain,
a Survey of Grave Treatment and Furnishing AD
43–410, BAR British Series 219, Oxford.
Pinter-Bellows, S., 2001, ‘Animal bone’, in Leach
2001, 289–303.
Pitt Rivers, A., 1884, Report on Excavations in the
Pen Pits near Penselwood, Somerset.
Price, J., 1980, ‘The Roman glass’, in G. Lambrick,
‘Excavations in Park Street, Towcester’, Northamptonshire Archaeology, 15, 35–118.
_____ , 1984, ‘The objects of glass’, in M. Fulford,
The Silchester Defences 1974–80, Britannia
Monograph 5,116–18.
_____ , 1985, ‘Early Roman vessel glass from burials in Tripolitania: a study of finds from Forte della
Vite and other sites now in the collections of the
National Museum of Antiquities in Tripoli’, in D.
Buck and D. Mattingly, Town and Country in Roman Tripolitania. Papers in Honour of Olwen
Hackett, BAR Int Ser, 274, 67–105.
_____ , 1987, ‘Glass from Felmongers-Harlow in Essex. A dated deposit of vessel glass found in an
Antonine pit’, Congrés AIHV, 10, 185–206.
_____ , 1989, ‘The Roman glass’, in S. Frere and J.
Wilkes, Strageath: Excavations within the Roman
Fort, 1973–1986, 192–203.
_____ , 1990, ‘Roman vessel and window glass’, in
McCarthy 1990, 163–79.
_____ , 1995, ‘The glass vessels’, in W. Manning, J.
Price and J. Webster, The Roman Small Finds.
Report on the excavations at Usk 1965–1976,
Cardiff, 139–91.
Price, J, and Cottam, S, 1994, ‘Glass’, in S. Cracknell
and C. Mahany, Roman Alcester: Southern Extramural Area. 1964–1966 Excavations, CBA Res
Rep 97, 224–9.
_____ , 2001, ‘The Roman glass’, in Leach 2001,
170–6.
Rahtz, P.A., 1977. ‘Late Roman cemeteries and
beyond’ in R. Reece (ed) Burial in the Roman
W 51
SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011
World, CBA Res Rep 22, 53–64.
Resnick, D., and Niwayama, G., 1988, Diagnosis of
Bone and Joint Disorders, 2nd edn, London.
Roe, F.E.S, 1995, ‘Stone’, in Coles and Minnitt 1995,
161–6.
_____ , 2000, ‘Worked stone’, in Barrett et al. 2000,
148.
_____ , 2001a, ‘Querns and millstones’, in Leach
2001, 235.
_____ , 2001b, ‘Whetstones’, in Leach 2001, 235–7.
_____ , 2001c, ‘Stone mortars’ in Leach 2001, 177–
8.
Rogers, G.B., 1974, Poteries Sigillées de la Gaule
Centrale, 1 Les motifs non figures, Sup Gallia 28.
Rogers, J., Waldron, T., Dieppe, P., and Watt, I., 1987,
‘Arthropathies in palaeopathology: the basis of
classification according to most probable cause’,
Journ of Archaeol Science, 14, 179–83.
Scheuer, J.L., Musgrave, J.H., and Evans, S.P., 1980,
‘The estimation of late fetal and perinatal age from
limb bone length by linear and logarithmic
regression’, Annuals of Human Biology, 7, 257–
65.
Stace, C., 1991, New flora of the British Isles,
Cambridge.
Stallibrass, S., 1991, A Comparison of the
Measurements of Romano-British Animal Bones
from Periods 3 and 5, Recovered from Excavations
at Annetwell Street, Carlisle, Ancient Monuments
Laboratory Report 133/91.
Stanfield, J.S., and Simpson, G., 1990, Les Potiers de
la Gaule Centrale, Gonfaron.
Steele, D.G., 1976, ‘The estimation of sex on the basis
of the talus and calcaneus’, American Journ of
Physical Anthropology, 45, 581–8.
Steinbock, R.T., 1976, Paleopathological Diagnosis
and Interpretation: Bone Disease in Ancient
Human Populations, Springfield, Ill.
Stevens, C., 1999. ‘The plant remains’, in Broomhead
1999, 156–65.
Stevenson, R.B.K., 1956, ‘Native bangles and Roman
W 52
glass’, Proc Soc Antiq Scotland, 88, 208–21.
_____ , 1976, ‘Romano-British glass bangles’,
Glasgow Archaeological Journal, 4, 45–54.
Stewart, T.D., 1979, Essentials of Forensic
Anthropology, Springfield, Ill.
Straker, V., ‘Charred plant macrofossils’, in Leach
with Evans 2001, 303–7.
Stuart-Macadam, P., 1985, ‘Porotic hyperostosis:
representative of a childhood condition’, American
Journ of Physical Anthropology, 66, 391–98.
Suchey, J.M., Brooks, S.T., and Katz, D., nd,
Instructions for use of the Suchey-Brooks system
for age determination of the female os pubis,
unpub report.
Sunter, N., and Woodward, P.J., 1987, Romano-British
Industries in Purbeck, Dorchester.
Thomas, C., 1987, ‘The worked stone’, in Sunter and
Woodward 1987, 30–6
Todd, M., 2007. Roman Mining in Somerset, Exeter.
Trotter, M., 1970, ‘Estimation of stature from intact
long bones’, in T. Stewart (ed) Personal
Identification in Mass Disasters, National
Museum of Natural History, Washington, DC, 71–
83.
Webster, C.J., and Croft, R.A., 1994. ‘Somerset
archaeology, 1993’, SANH 137, 127–56.
Wedlake, W.J., 1958, Excavations at Camerton,
Somerset, Bath.
Wheeler, R.E.M, 1929, ‘A Roman pipe burial from
Caerleon, Monmouthshire’, Antiquaries Journ, 9,
1–7.
White, K.D., 1970, Roman Farming, London.
White, R.W., 2007. Britannia Prima: The Romans in
the West of Britain, London.
Wilson, E.M., 1980, ‘Excavations at West Mains of
Ethie’, Proc Soc Antiq Scotland, 110, 114–21.
Woodward, A.B., 1993, Coffins and grave linings, in
Farwell and Molleson 1993, 227.
Workshop of European Anthropologists, 1980,
‘Recommendation for age and sex determination’,
Journ of Human Evolution, 9, 517–49.