An Archaeology of Brief Time: monuments and seasonality in Roman Britain

by Raphael MJ. Isserlin

Introduction
It is a fundamental tenet of Romano-British studies that inscriptions with Imperial or other
official titles can date buildings. But to what does one assign them – the calendar year or the
season? Archaeologists’ perceptions of time, by and large, use linear (or institutional) time
(Lewis & Weigart 1990: 84), and cyclical’1ime (‘a dance on one Gurvitch 1990: 71) only
rarely. I would like to explore whether it is possible to relate activity at a site to a particular
time of year, for I feel that recent work (Barrett 1993) has barely scratched the surface of this
problem.
Seasonality and inscriptions
One specific example, arising from the building programme of Gn. luI. Agricola, demonstrates
how vital these concepts are. It is often assumed that the Vemlamium fomm-inscription can be
dated to a particular time of year within Agricola’s governorship. Certainly we are wellinformed of this governor thanks to Tacitus, and to archaeology (though some doubt this:
Hanson I 987). But closer scmtiny reveals no little disagreement. The inscription has been
dated to autumn AD. 79 (Frere 1983: 55), winter AD. 79 (Frere 1987 99), or even AD. 81
(Eck 1970: 48-9). The reason for the inconsistency is simple: the text preserves no seasonal
detailsl The most recent verdict, based on institutional time, assigns it to July – December
AD. 79 (Tomlin 1993: 137). Details of Agricola’s governorship are cross-linked to the broader
chronology of the regnal years of the Emperor. But the seasonal argument rests on purely
external sources (Agricola: 21) and the premise that autumn or winter can be unfavourable to
outdoor work (building, military campaigning and, all too often, archaeological excavation)!
Possibly the governor (Gn. luI. Agricola) returned from campaigning in northern England
to preside over an opening ceremony in the southern civil zone. The Gaulish ara provinciae at
LyoniCondate was dedicated on the 1st August, 12 B.C. (an auspicious date in the Gaulish and
Roman calendars: Wightman 1985: 51). It has been suggested that Trier’s ‘birthday’ was on
this date (Heinen 1984: 41).
The one Romano-British ‘historical’ example of seasonal monument-dedication lacks hard
epigraphic proof and rests on circumstantial textual evidence. But a more fundamental question
is clear. Were there times when Romano-British monuments could be dedicated – or offerings
made at them? Above all, does evidence of cyclical time exist? The idea that seasonal festivals
migl1t have been held in the province was put before a wider public from literary evidence and
analogy (Henig 1982). Some Romano-British villa mosaics depict the seasons (Ling 1983)
There is clear epigraphic evidence for seasonality at monuments (Table 5. I). 46 Raphael M.J. [sserlin
ROMAN DATE MODERN DATE YEAR FESTIVAL COMMENTS RIB
KALlAN 1st January 241 new year altar 1983
DRS 21 st April ? Parilia altar 1270
AOKALIUN … May 278 .).) beltane tomb 1255
(A 0 ) XI KAL SEPIEM 22nd August 262/6 hetween consualia & dedication slab 605
volcanalia
ADXKSEPI 23rd August 152 I month before statue 309
volcat/alia
vrnr KAL OCTOB 23rd September 244 autumn equinox; door jamb 327
VlII A DK OC 23rd September 234 autunm equinox: dedication slab 328
mercalus
(A D) Xlll [KALI ET XIII 19th. 20th October 241 armiluslriuTII altar 882
KALNOV
(A D) xm [KALI ET 19th, 20th October 24 112 armiluslrillTII altar 883
(A D) xm KAL NOV
(A 0) III KAL NOVEM 30th October 221 day before samuim dedication slab 1465
Table 5.1: Seasonality and Festivals as reflected in inscriptions on British monuments
[Celtic festivals: bold type; Roman festivals Italics]
(Sources: RlB L Burl 1983: 34; Ross 1967: 83; Scullard 1981)
Seven out of ten inscriptions fall within a three-month period; only two are definitely not
3rd-century. Particularly significant according to the Roman (Julian, solar) Calendar was the
1st January (New Year’s Day: RJB 1983), when new consuls were appointed, temples dedicated
(Scullard 1981: 52ft’.), gifts given, and, from Augustus onwards, statues erected in and out of
Rome (Millar 1992: 142–‘3). On that day the Roman army renewed its vows, set up new altars
and buried old ones (the Maryport altars indicate this: Wenham 1939); in A.D. 69, the army
tluew stones at, and demolished, statues (Tacitus: His!., I, 55). One inscription (DRS: Day
Sacred to Rome) is Rome’s Official Birthday – an occasion so obvious, it did not need stating
that it was the 21st April (RJB 1270; Risingham; Scullard 1981 : 101 ff). The volcanalia was a
festival dedicated to Vulcan in August, and had to be treated with particular care; another
festival, the consualia fell two days earlier and RJB 605 was dedicated in between these dates
(Scullard 1981: I 78ff.). Other astronomically-derived dates were significant to an agrarian
society (23rd September, the autumn equinox; the date appears twice, on one site, ten years
apart – RIB 327 & 328). It was also the end of the 3-day festival of mercalus, when fairs could
be held (Scullard 1981: 188). Most auspiciously, it was the emperor Augustus’ Birthday –
surely commemorated by the Leg. II Augusta. October saw the festival of armilustrium, when
weapons were ritually purified (Scullard 1981: 197-8). One date is very close to the date of a
Celtic festival (Samuin is actually on the 1st of November: RJB 1465)
No doubt it was important to inaugurate a monument at the right time, and here there
appears to have been a synthesis of things Roman and Celtic. The significance of others is at
present unclear (RJB 882 and 883), though we can guess that some monuments were
inaugurated a month before the supposed date of the equinox (RJB 605; 309) This could
indicate a lunar calendar operating, but the dates being commemorated in Roman fashion, i.e., An Archaeology of Brief Time: monuments and seasonality in Roman Britain 47
cyclical and institutional times in conjunction. Operating with two calendars is known in
ludaea only 22 years after the incorporation of the province in A.D. 106 (a marriage contract
from the Babatha archive uses the Latin and the Greek calendar dates, but not the Hebrew: P.
‘ladin Lewis 1989). There is no reason why the North-Western provinces should have
operated with only one chronology.
A more general trend?
Concern to do the right thing at the right time appears have been particularly prominent in the
3rd century. But is it a phenomenon confined specifically to Britain or to the 3rd century? The
most obvious way to check this is to examine the material from the neighbouring provinces of
Gallia Belgica and Germany, to be found in CIL XIII (Table 5.2)
ROMAN DATE
PRIKAL MAl
VID MAl
DDIDKALTUN
I .. IDUS [fU]LlAS
VII K(AlENDAS)
SEPTEM[BRESI
VIII K OCTOB
IDNONOCT
MODERN DATE
29 April
II May
30 May
13 July
24
23 September
5 October
YEAR
198
?
232
245
44
124
FESTIVAL COMMENTS CIL XIIlI No
/udijlorae building slab 4132
altar 4679
huli apollillares building slab 4131
mundus palel brooze tablet 3498
autunm equinox: building slab 4565
merealus
lIIulldus palel building slab 4149
Table 5.2.· Seasonalily andlestivals as rej1ected in inscriplions on Gm/fish Nfonuments
[Celtic festivals: bold Roman festivals: Italics]
(Sources: ClL XIII; Burl 1983: 34; Ross 1967: 83; Scullard 1981)
One item (not discussed further) falls outside the bulk of the pattern observed so far – CIL
XIIJ390, If If id marlii – and so it should, for on purely epigraphic grounds it may be f31sum.
But on the whole, the range is tight. A bronze tablet, states the constitution of a collegium,
from near Amiens (CIL XIII. 3498) including an ‘annual general meeting’. Such a fixed event
is precisely what one might expect of behaviour-patterns associated with monuments –
cyclical activity. The date, 24th August, is one of two days in the year when the sacred shaft in
Rome was opened – the mundus (Scullard 1981: 180–1). The other was the 5th of October
(Scullard 1980: 180, 191) – represented by CIL XlII.4149 from lunkerath. It commemorates
tile impermanent action as well as stating the nature of the permanent monument – a building
was dedicatum. Not surprisingly, the building was a temp/um! The collegia, mentioned in the
bronze tablet from Anliens, probably met in such a building. It would not be surprising to find a
ritual shaft anywhere in Picardy. One of the main features of deposits in ritual shafts is the
repetitive – almost cyclical – structuring of their deposits (see below pages 50-53)
A building-inscription from Bitburg ushers in the 3rd-century (CIL XIII 4132). It says a
tr’ibunal and proscenium were erected and ludi (games) were given. It falls on the date of the
ludi florae (Scullard 1980: 270), related to fertility. Another set of games was the ludi
Apollinares (Scullard 1980: 164), and their date is mentioned on a rebuilding inscription, 31so 48 Raphael M.J. Isserlin
from Bitburg (ClL XIII.413 I ). This has the name of the emperor Philip erased, and
commemorates the rebuilding of a farator. Presumably this was a particularly harmonious day
to hold such an event, with the games part of the opening-ceremony.
Of the 2314 inscriptions in RIB 1, ten are inscribed thus – under 0.004%. Of those 1487
from Gallia Belgica in CIL XIII 2.1 (nos. 3253-4740) only seven – under 0.005% are thus
inscribed. There are apparently none in the latest volumes of R.l.G. The percentages – if the
low numbers have any meaning – are almost identical. But what do they mean as social
indicators? What lies behind these expressions of concern – such cuckoo-clock timing? For
there is a clear desire to mark time far more in the 3rd century than there was before, when the
seasonal dates appear only sporadically. Equally, why was this not so much the practice in
earlier times?
In a way, it was. But not so publicly, for it was, initially, performance of the deed, not the
display of the time it was done at, that mattered. Such things were governed by long-standing
tradition. Our knowledge of Gaulish festivals derives from the calendars of Villards d’Heria
and above all, CoJigny, bronze plaques set up on a wall of some monument to advise of what
the right times to do things was. These were designed to be displayed to a literate audience. The
latest research suggests that according to the style of lettering employed, Coligny may be 2nd
century or later. Gaulish (or Celtic) time-keeping, based on a lunar calendar, is recorded here, a
century and a half after the Julian Calendar reforms (Duval & Pinaut 1986 35). Tradition died
hard. Some festivals were shifted – the floralia from 1st May to 30th April (CIL III:
Wightman 1985: 187). British timekeeping may have operated in the later Iron Age on an
astronomically-derived system too: Diodorus Siculus (50–30 BC) preserves a passage of
Hecateus, on a tradition that, some think, refers to Britain:
It is said that the god [Apollo] returns to the island every 19 years, the period
in which the return of the stars to the same place in the heavens is
accomplished
(Histories: V).
Did a lunar calendar operate in the Roman period? How did it relate to the Gaulish one?
Checking through the other volumes of CIL, or IGRR, may reveal other patterns, or the
same low numbers of such inscriptions. Conjunctures of time and action in other provinces
depended on non-local calendars. An altar from Lectoure in Gaul depended on a non-Gaulish,
and non-Roman calendar (CIL XlII 525). On March 24, 239 (the 3rd-century again!)
a lady ‘received the potency’ (vires) of one Eutyches. The words described
emasculation, and as the day was the ‘Day of Blood’ in the cult of Attis, the
occasion was presumably the castration of Eutyches, a priest in the service of
the god. The woman received the pieces
(Lane Fox 1986: 348).
The timing is the cyclical timing of an Eastern religion, in a Western province. A ritual
castration-clamp used in the cult of Cybele was recovered from the Thames (francis 1926).
Castrations were probably carried out in Londinium (it ‘shows signs of much usage and had
been carefully repaired in Roman times’ (Marsden 1980: 137)) and probably at specified times. An Archaeology of Brief Time: monuments and seasonality in Roman Britain 49
Identifying seasonality through archaeology
We may summarise the information from these inscriptions by arranging the dates sequentially.
A cumulative diagram shows the range (Table 5.3).
lB
Jan Feb
? B
2lB IIG
29G 30G
Mar Apr May Jun
23B 23B 19/20B
23B 23B 19120B
13G 24G 23G 30B
Jul Aug Sep Oct Nov Dec
Table 5.3: Cumulative patterns of monument dedication, in Britain and Gallia Belgica, sorted
by date and province
(Sources RIB 1; elL XIII) [B=Britannia; G= Gallia Belgica/Germania].
Some times of year seem to be avoided, and excavated environmental evidence too may reflect
this: were sites used all year round, or only at specific times of year? A glance at works dealing
with many ritual shafts for instance, suggests that they yield evidence for season-specific
activity: nuts, berries, etc. seem to suggest autumnal activity. Twigs with evidence of budscale
(or not!) could be equally informative (perhaps it is worth looking at the vast collection of exvotos in wood from the source of the Seine once more; most of them are twigs or branches).
And there are the animals for sacrifices, represented by animal bones. Do we reasonably infer
springtime sacrifices of lambs from the immature ovicaprids? And what about that matter of
‘autumn killing’ – does that only occur in agricultural contexts? A glance at the votive bonedeposit from Uley suggests not – perhaps 80% of the sheep and goats found, were killed, some
150 each month, over the autumn months, every year for 50 years (Levitan 1993:
Woodward & Leach 1993 300-1), and the rest perhaps on the Ides of every month – the
practice of ovis idulis (Frayn 1984 39). The same pattern may hold good for the sacrifices at
Harlow (Legge & Dorrington 1985: 132). The Vindolanda tablets record sacrifice on the dies
Kalendarum – a term often used without specifying a particular month to refer to New Years
Day (Tab. Vindol. II 265). They also mention at least one summer-based festival (Tab. Vindol.
II 190; Bowman 1994: 115- 7) involving the use of wine for a goddess (which one?) on or
around the 25th June, a date absent from our inscriptions but within this seasonal span.
With this in mind, let us take two published examples – Jordan Hill, near Weymouth,
Dorset, and Ashill, Norfolk. Material was deposited during the 3rdl4th century in a ritual shaft
in the temple courtyard at Jordan Hill. After the shaft was dug, a stone cist was built at its base
around iron objects and pots. This primary deposit was sealed by a succession of 16 sets of
deposits (Table 4). Each set of deposits commenced with a layer of stone slabs, on top of which
ash, containing bird-skeletons, was dumped. The deposit was sealed by a further stone slab, on
top of which a coin and a bird was placed. Subsequently another set of deposits would be
dumped (consisting of slab, ash, bone, slab, and a further coin and bird). When the shaft was
half-full, another cist was constructed containing a sword, spearhead and pots. Thereafter the
sequence of slabs and dumping resumed. 50
Stone
slabs
Animal bones
raven
crow
buzzard
starling
L
I
N
E
A
R
T
I
M
E
Raphael M.l. Isserlin
Serving-vessels
pottery
Iron
sword
spear-head
knife
steelyard
iron bar-fragments
Bl”Onze
500 coins
Other organic
(Wood?) – ash
Table 5.4: The Jordan Hill deposit, sorted by category
(Sources: Drew 1931; Wait 1985: 328, 410)
CYCLICAL TIME
Fig 5.1. The relationship between Linear and Cyclical time as they afJect the accumulation of
deposits, displayed on a 2-dimensional graph. (.S’ource: author). An Archaeology of BriifTime: monuments and seasonality in Roman Britain 51
The structured nature of the artefact-deposition is readily apparent. Equally clear is the regular,
cyclical, activity reflected in the stratigraphic sequence. It may reflect a ritual purpose such as
sacrifice or cleaning of the temple it surely reflects cyclical time. The construction of
the second cist, half-way down the shaft, may reflect some process of reinvigoration
(rededication?) perhaps connected to a rebuilding of the main temple building. Almost 5,000
more coins and other finds have come from other dry-land ‘hoards’ from Jordan Hill (Barrett
1909; O’Neill 1936; Salisbury 1930). Only full re-examination of this material will help set the
timing of the depositions of these sets of material in the ritual shaft in a broader chronological
context. Here is another example of cyclical time, here mingled with the linear time of the
stratigraphic record. The interrelationship is displayed in Fig 5.1.
The other example comes from within the (lst century) square enclosure at AshilI, Norfolk,
0 PENANNULAR
M
I 6 0 5
m
– BASKET
I
10 0 S KNIFE. WALL’ PLASTER &
WHETSTONE
15 0 – BUCKET & FOUR SANDALS
,
i 19
., ,.
– BROOCH & LATCH-LIFTER
I ” 21,
(6
28 ” –
BASKET
,
30
32
33 AMPHORA. BUCKET & QUERN
31, KNIFE-HANDLE
I 36 BASKET
,i
38
II 40
• URN 0 SHERDS i FLAGON OAK PLANK
@ ‘URNS ‘ 5 SAMIAN WARE “, ,
LEAVES etc
Fig 5.2. Reconstructed schematic section ofshaft no. 3 at Ashill, Norfolk
(oS’ouree: Gregory 1977: fig. 5) 52 Raphael M.J. Isserlin
excavated in 1874 and 1961 (Gregory 1977). Opinion is divided as to whether or not it
contained ritual shafts or merely wells; it has certainly been suggested that the site is a
Viereckschanze, so a ritual interpretation is a possibility for the deposits in these features. The
(reconstructed) sequence of deposition of fills and artefacts for shaft no. 3 is given in Fig 5.2.
The primary deposit included a haunch of venison; thereafter a sequence of 11 deposits of
pottery vessels embedded within apparently separate, distinctive organic fills, including
material from hazel trees. At least 50 complete vessels were deposited, occasionally within
basketry (Table 5.5).
Animal bone
deer
pig
goat
cattle
frog
toad
Stone
whetstone
Iron
bucket
knife
latch-lifter
Other organic
leaves
twigs
nuts
Copper alloy
brooches x2
Pottery
50–1000) vessels
including 19 jars,
4 bowls, 3 flagons,
1 urn & 1 amphora
Leather
sandals x4
Table 5.5: The Ashill deposit, sorted by category (Source. Gregory 1977).
The original accounts make it clear that the vessels were deposited with some care (Gregory
1977: l3ff. for details) It is clear that there is some sort of cyclical repetition of elements,
though not so well defined as at Jordan Hill. Activity was carried out in autumn, for,
it is worthy of remark that the nuts were in different stages of growth; those at
the bottom had nuts just formed, whilst a few feet higher they were more
matured; and higher still they were perfect
(Barton 1879: 228).
One modern opinion is that these deposits were deposited ‘between late spring and autumn of a
single year at intervals of several weeks’ (Gregory 1977: 13) – in effect. an extended single
deposit. If however the shaft were covered and uncovered (ie., on the festivals of mundus patel)
then no deposits would accumulate while the shaft was closed and the various strata would
develop separately. In that case it would be possible to envisage successive, cyclical deposits in
Autumn – perhaps in August or October (Table 2 for festival dates)
There is plenty of organic material to go at there’ It might be, with modern techniques of
excavation, that environmental material might be recovered, in greater quantities, and be rather
more susceptible along the lines of analysis indicated above. There are problems and, in the
case of ritual shafts, quite literally pitfalls. Nevertheless, it would be nice to see what results a
judicious and sympathetic programme of sampling of organic material migllt bring. Alas, we An Archaeology of Brief Time: monuments and seasonality in Roman Britain 53
lack super-accurate dating – to calibrate rates of seasonal accumulation within the
microstratigraphy. Though we may begin to understand context-formation processes, giving
them a linear chronology eludes us still.
Nor was the Roman period the only occasion when changing attitudes to time affected the
fabric of communal religious monuments. In English churches, bells were erected during the
12th-14th centuries (Thrift 1988: 66), while dates were increasingly specified in documents; in
public places such as markets and the 14th century court, timekeeping devices appeared (Thrift
1988: 74). They marked out units of labour (as urbanisation and industry increased) for the
convenience of those at the top, the managers (Thrift 1988: 78-82). How very different from
the increasingly over-bureaucratised, and privatised top-heavy society of the Later Roman
Empire! Something like time-and-motion-studies were undertaken at the jabricae [arms
factories] of Egypt, Antioch and Constantinople. [n AD. 374 a specific production-rate per
worker per month of helmet shells and inlaid decoration was noted at Antioch, and an identical
quota demanded of the workers at Constantinople (Jones 1964 835). In Egypt a 2ndl3rd
century papyrus records the output of just two days (Bruckner & Marichal 1979: no. 49)! In
Britain various tile-graffiti record production-rates too (not all necessarily 3rd century,
however: I hope to return to the topic of production another time).
Twenty years before the management of the fabricae was put in hand, someone wished to
codify and communicate festivals in a way not practised for several centuries. The result was
the Codex-Calendar of AD. 354 (Salzman 1993) – not the first of its kind, but quite possibly
one of the last to mention pagan ones. British (and other) inscriptions discussed in the first part
of this paper are merely part of a trend. They are statements of meticulous observance, and
more explicit, with regard to timing, than anything that had gone before. Stating a mere regnal
year no longer sufficed. These were proclamations of serenity, in the midst of the ‘3rd century
crisis’. The very denial speaks volumes. Were they statements of personal faith, routine acts of
ritual – or symptoms of religious revival? In Rome, mentions of the Arval brethren become
increasingly more detailed at this time (Beard 1985 13lfC Scheid 1990: see Smith: this
volume, p. 60) Trier’s birthday seems to have been instituted in the late 3rd century
(Wightman 1985: 235). The same trend towards a precise relation of deeds can be seen in
rebuilding inscriptions. They often date, a recent Empire-wide survey has shown, to that same
3rd century (Thomas & Witschel 1992). They portray spurious or vastly over-exaggerated
major reconstructions. Such perfection of planning is a symptom of decay through
centralization, paralysis and intrigue (Northcote Parkinson 1986: 77, 86ft’.). Perhaps these
reflect the belief of individuals that they needed to validate their positions through display.
Belief was, indeed, the key. One interpretation of part of Book II of St Augustine’s
(‘onfessions is that he broke the circle of Roman time – the same sort of time as we have seen
in Diodorus Siculus – and replaced it with linear time (Hassard 1990: II) Whether or not one
chooses to believe this reading, in the smashing and burial of the bronze calendar at Coligny
this was almost literally the case. Critics agree that it was deliberately shattered before it was
buried with a statue of Mars, also broken (cf Duval & Pinault 1986: 37). If one were to give it
an historical context, it would be to see this as more than an ordinary rite of termination, for the
new order was to use a Christian calendar. There was no need of such things as pagan
calendars, except for their decorative motifs. 54 Raphael M.J. Isserlin
Postscript
Thjs paper arose out of a desire to arrange a limited amount of information along certain lines,
and see if any pattern resulted. Undoubtedly, the methodology adopted here is open to two
major criticisms – neither made at the Conference. Constructing festivals from dates on
inscriptions is risky, for there were so many Roman festivals – to resume the analogy with
mediaeval England, every day was a Saint’s day or festival (Hutton 1994). But this is to miss
the point – it was a date which was important to some people to cause an action to be
performed – be it the Saint’s namesake causing a Mass to be said, or a Roman who put up a
monument causing the inscription to be cut. It therefore is designed to let performers make a
point to an audience, to show that they were doing the right thing at the right time, building
monuments in the summer-autumn months, and making major sacrifices at them in the late
autumn and other ones in between.
The other major criticism concerns the size of the sample – under 1 % of inscriptions
Why are the other 99% of inscriptions ‘date-less’? I can offer no answer to that at present.
Trier’s birthday may have been shared by other cities. It is possible that this date, Rome’s
birthday, that of the reigning emperor or perhaps the Augustus’ birthday is when many
inscriptions were cut and monuments erected. But that is to propose another myth, based on no
real proof – just like the one at the beginning ofthis paper.
A cknowiedgemellts
Chief thanks go to my parents for encouragement and support at a difficult time. I am most
grateful to my supervisor, Mr. Brian Hartley (School of Classics, Leeds University), to Mr.
Mark Hassall (Institute of Archaeology, University College London), and to Mr. Colin Wallace
(Archaeology Section, Essex County Council) for reading and commenting on the text at very
short notice, and to Dr. Martin Henig (Institute of Archaeology, University of Oxford) for
refereeing and commenting on the text. I am also indebted to Dr. Christopher Smith (School of
Classics, St Andrews University) for references to the Arval Brethren of which I would
otherwise have remained ignorant limits of time mean that I have not been able to give
Scheid’s study the attention it deserves. These commentators are not responsible for the errors
that this paper contains Fig 5. I has been specially drawn by Stuart McNeill: Fig 5.2 is
reproduced from Gregory I 977. An Archaeology of Brief Time: monuments and seasOIUllity in Roman Britain 55
Bibliography
Barrett, J.e. 1993. Chronologies of remembrance: the interpretation of some Roman inscriptions, World
Archaeology, 25 no 2 (1993): 236-247.
Barrett, W.B. 1909. Preston, Dorset. Roman Antiquities, Notes and Queries for Somerset and Dorset 11:
64-66.
Barton, T 1879. Roman Discoveries at Ashill, Norfolk Archaeology VIII: 224-30.
Beard, M. 1985. Writing and ritual: a study of diversity and expansion in the Arval Acta, Papers of the
Bntish School at Rome, 53: 114-162.
Bowman, A 1994. Life and Letters on Ihe Roman Frontier.’ Vindolanda and ils people. (London).
Bowman, A & Thomas, J.D., 1994. The Vindolanda Writing Tablets (fabulae Vindolandenses If).
(London).
Bruckner, A & Marichal, R. (eds) 1979. Chartae Lalinae Anliquiores, Part X, Germany I. (Zurich).
Burl, A 1983. Prehisloric Astronomy and Ritual (Aylesbwy).
Drew, C.D. 1931. The Excavations at Jordan Hill, Weymouth, 1931, Proceedings of the Dorset Natural
Hislmy and Archaeological Society LID: 265-76.
Duval, P-M. & Pinault, G. 1986. Receuil des InSCriptions Gauloises (R.J.G.), Volume Ill: Les
Calendrie,.s. (Coligny, Villards d ‘Heria). (XL Ve supplement a Gallia). (Paris).
Eck, W. 1970. Senatoren von Vespasian bis Hadrian. Vestigia Xlll: 48-9.
Francis, AG. 1926. On a Romano-British Castration Clamp used in the Rites of Cybele, Proceedings of
Ihe Royal Sociely ofll/fedicine XIX: 95- 110.
Frayn. J.M. 1984. Sheep Rearing and the wool trade in Italy during the Roman period. (Liverpool).
Frere, S. S. 1987. Brilannia. (thi.rd edition). (London).
Frere, S.S. 1983. Vent/amium Excavations, [J. Research Report of the Society of Antiquaries of London,
XXXXI (London).
Gregory, T 1977. The Enclosure at Ashill. U1 Wade-Maltins, P. (ed) Norfolk, 9-30. East Anglian
Archaeology Report No.5.
Gurvitch. D. 1990. Varieties of Social-ti.me. iJ1 Hassard, J. (ed) The Sociology o/Time, 67-76.
Hanson, W.S. 1987. Agtlcola and Ihe Conquest o/the North.
Hassard, 1. 1990 Introduction: The Sociological Study of Time. In Hassard, 1. (ed.) The Sociology of
Time. 1- 18.
Hassard. J. (ed.) 1990. The Sociology of Time.
Heinen. H 1984. Augustus in Gallien lmd die Anfange des romischen Trier. In Rheinisches
Landesmuseum Trier, Trier, Augustusstadt der Treverer: Stadt lind Land in vor- lind
/riihromischer Zeil, 32-46. (Mainz).
Henig, M. 1982. Seasonal feasts in Roman Britain, Oxford JOllrnal of Archaeology 1: 213-223.
Hutton, R. 1994. The Rise and Fall 0/ Merry England. (Oxford).
Jones. A.H.M. 1964. The Laler Roman Empire, 284-602: A SOCial, Economic and Administrative Survey.
(Oxford).
Legge, A & Dorrington, AJ. 1985 The Ani.mal Bones. In France, N. E. & Gobel, B. The Romano-British
Temple at Harlow, 122- 132. (West Essex Archaeological Group). (Harlow).
Levitan, B. 1993. Vertebrate Remains. In Woodward. A & Leach, P. (eds) The Uley Shrines: Excavation
of a ritual complex on West Hill. Uley, Gloucestershire: J 977-9, 257-301. English Heritage
Archaeological Repolt No. 17 (London).
Lewis, J.D. & Weigalt, AJ., 1990. The Structures and Meanings of Social-ti.me. In Hassard (ed) The
SOCiology o.fTime, 77- 101.
Lewis, N. (ed) 1989. Judean Desert Studies: The Documents from the Bar-Kochba Period in the Cave 0.( 56 Raphael M.J. lsserlin
Letters: Greek Papyri. (Israel Exploration Society. Hebrew University of Jerusalem)
(Jerusalem). [= P. Yadin].
Ling, R.1. 1983. The Seasons in Romano-British Mosaic Pavements, Britannia XIV: 13-22.
Marsden, P. 1980. Roman London (London).
Millar, F. 1992. The Emperor in the Roman World (second edition). (London).
Northcote Parkinson, C. 1986. Parkinson’s Law, or the Pursuit 0/ Progress. (The Library of Management
Classics). (London).
O’Neill, B.H. St.J. 1936. Coins from Jordan Hill Roman Temple, Proceedings of the Dorset
Archaeological Society 57: 140-142.
Ross, A. 1967. Pagan Celtic Britain (London).
SalisbulY, F.S. 1930. A hoard of Roman coins from Jordan Hill, Weymouth, Proceedings 0/ the Dorset
Archaeological Society, 51: 158-182.
Salzman, M. R. 1993. On Roman Time: The Codex-Calendar of 35-1 and the Rhythms of (/rban L!fe in
Lale Antiquity (The Transformation of the Classical Heritage). (Berkeley, Los Angeles &
Oxford).
Scheid, J. 1990. Romulus et ses (reres.· Le College des Freres Arvales, modele de culte Public dans la
Rome des Empereurs. Bibliotheque des ecoles fi’anyaises d’ Athenes et de Rome, fascicule 275
(Rome).
Scullard, H H. 198 1. Fest ivC/ls and Ceremonies of the Roman Republic (London).
lllOmas, E. & Witschel. C., 1992. Constructing reconstruction: claim and reality of Roman rebuilding
inscriptions from the Latin West. Papers oflhe Hritish School at Rome LX: 135-177.
Thrift, N. 1988. Vivos Voco: Ringing the Changes in the Historical Geography of Time Consciousness. in
YOWlg, M & Schuller, T (edl’) The Hhythms o/Society, 53-94. (London).
Tomlin, R.S.O. 1993. Roman Towns and Roman Inscriptions of Britain, 1939-89. (n Greep, S. (ed.)
Roman Towns: the Wheeler inheritance, A review of 50 yean’ research, 134-146. Research
Repolt No. 93. Council for British Archaeology (York).
Wait, GA. 1985. Religion in Iron /Jge Britain. British Archaeological Reports 149 (Oxford).
Wenham, L.P. 1939. Notes on the Ganisoning of Malypolt, Transactions of the Cumberland &
rYestmorland Antiquarian and Archaeological SocietY’ (2nd Series), XXXIX: 19-30.
Wightman, E. J985. Callia Belf{ica (London).
Woodward, A. & Leach, P. 1993. Synthesis. In Woodward, A & Leach, P. (eds) The Uley Shrines:
Excavation of’a rill”,1 complex (In ff’esl 11,11, [I/ey, Cilollcestershire: 1977-9, 303-335. English
Heritage Archaeological Repol1 No. 17 (London).
Woodward, A. & Leach, P. (eds) 1993. The [!ley Shrines. Excavalion of a rillial complex on West Hill,
Uley, Gloucestershire. 1977-9. English Heritage Archaeological Report No. 17 (London).