Excavations at Terryhoogan , Co Armagh: Early Medieval house

Ulster Archaeological Society

EXCAVATIONS AT TERRYHOOGAN, COUNTY ARMAGHAuthor(s): CORMAC McSPARRON, PHILIP MACDONALD and GILL PLUNKETTSource: Ulster Journal of Archaeology, Third Series, Vol. 66 (2007), pp. 120-131Published by: Ulster Archaeological SocietyStable URL: http://www.jstor.org/stable/20699260 .

Accessed: 29/09/2014 13:39

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

Ulster Archaeological Society is collaborating with JSTOR to digitize, preserve and extend access to UlsterJournal of Archaeology.

http://www.jstor.org

This content downloaded from 143.117.16.36 on Mon, 29 Sep 2014 13:39:24 PMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=uas

http://www.jstor.org/stable/20699260?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp


Ulster Journal of Archaeology, Vol 66, 2007

EXCAVATIONS AT TERRYHOOG AN, COUNTY ARMAGH CORMAC McSPARRON

Centre for Archaeological Fieldwork, School of Geography, Archaeology and Palaeoecology, Queen's University, Belfast

c mcsparron @ qub .ac .uk

with contributions by PHILIP MACDONALD and GILL PLUNKETT

Excavations at Terryhoogan, Co Armagh, in 2004 uncovered a circular gully cut into subsoil with a

diameter of 9.4m. This is interpreted as the drip channel and drain of a round house of the Early Christian period. A linear ditch, dating to the 13th or 14th century, seems likely to have been a boundary marker.

INTRODUCTION An excavation was undertaken from 13 April to 28

May 2004 at Terryhoogan, Scarva, Co Armagh, by the Centre for Archaeological Field work, on behalf of the Environment and Heritage Service: Built

Heritage. The site at Terryhoogan was under

immediate threat of destruction due to the proposed construction of a single dwelling. Test trenching by Stiof?n ? Cathmhaoil of the Environment and

Heritage Service: Built Heritage, suggested the

potential of the site, and in the light of this, the Environment and Heritage Service: Built Heritage decided to support an excavation. The aims of the

excavation were to preserve the site by record, to

identify its nature, function and date and place it within its archaeological context.

The excavation uncovered a circular gully,

possibly a drainage gully around an Early Christian

Down R.Lagan

Arm?gh

Armagh ^ Scarva

Fig 1 Site location.

round house, with two curving extension gullies,

possibly part of a drainage gully around an annexe to the round house. There was no trace of a structure

within the gully but there had been considerable truncation and disturbance of the site due to

ploughing and the digging of field drains. The

drainage gully appears to have filled up rapidly and

possibly to have been cleaned out on one or more

occasions. At the south-east the gully was somewhat

wider and stone packed. Considerable amounts of

souterrain ware pottery came from this area and it is

likely that this combination of stone packing and a

large number of finds indicates the presence of a

causeway through the gully and an entrance to the house.

The site is located on gently sloping ground some lkm from Scarva village (IGR J3054 3443) at an altitude of approximately 30m OD (Figs 1-3). The

current land use around the site is mostly pasture with little tree cover.

There has been ecclesiastical activity in the

locality of Terryhoogan since the Early ?Christian period. The neighbouring townland of

Aughlish contains the ancient church and

graveyard of Ballynaback (Arm 14:10), located some 200m south-east of the development site

(Fig 2). The graveyard is approximately oval, measuring 70m by 46m. In its centre there is a raised area which is likely to be the site of a church. Ballynaback is equated by Muhr (2001, 304) with the early Tamhlacht Gliadh, a site referred to as Hacyglid in the 1306 Taxation of

Pope Nicholas (Reeves 1847,253). The second ancient graveyard in the area is at Relicarn (Arm 14:09) in Terryhoogan, located some 150m south-south-west of the site (Fig 2). This

- graveyard is also oval, measuring 66m by 46m, and is surrounded by a low bank and hedge. It

120



I 7 014:018 I ̂

I t^^^ ['

^^^^ Excavation site ji ^ \ j

/ \v /7 l?

' Scarva

c /7,;' "

iL ) x\ //

?"" \

f ? 2004 Excavation site , 014:009 - Relicarn Graveyard

Jk 0 Archaeological site 014:010 - Ballynaback Graveyard !-. 014:011 - Rath

'-' LaKe 014:018-Crannog?

ZZ Road - Townland border

County border ? 500 metres

Water channel

Fig 2 Archaeological sites in the vicinity of Terryhoogan.

is suggested by Hamlin (1976, 533) that the

graveyard may be quite late, although it was the

find-spot of a bell known as the Bell of Armagh. This bell bears an inscription which asks for a prayer for Cumascach, son of Ailill, who can be identified as a steward of Armagh who died in AD 909 (ibid). A small walled grave is said to be the burial place of Redmond O'Hanlon (Day & McWilliams 1990,19), the famous 17th-century raparee.

There is a destroyed rath (Arm 14:011) about lkm south-east of the site of the excavation, and a

possible crannog (Arm 14:018) some 850m to the north-west (Fig 2).

METHODOLOGY

Although there was some removal of

sod and upper topsoil by mechanical excavator, under archaeological

supervision, excavation of the lower

topsoil and archaeological strata was

carried out by hand. All excavated

archaeological features and layers were

recorded using the context recording

system supplemented with scale

drawings and photographs. Recording was carried out in accordance with the

techniques set out in the School of

Archaeology and Palaeoecology's Excavation Manual (2003) and the Environment and Heritage Service's

Excavation Standards Manual (2001).

THE EXCAVATION An area of approximately 16m

(east/west) by 20.5m (north/south) was

mechanically stripped and manually cleaned to the top of the archaeological features.

Upon removal of the brown loam

topsoil a number of Early Christian features, beneath later agricultural

drainage features were apparent, cut into

the orange clay subsoil. A series of cultivation features had to be removed

to fully expose the archaeological features, especially in the western half of the site (Figs 4, 5).

The primary feature was a large circular gully (C3), approximately 0.70m wide and up to 0.40m deep with a diameter averaging 8.0m internally and 9.40m externally.

The circular gully (C3) was divided into eight sectors numbered 1 to 8 (Fig 4).

Archaeologically secure finds and samples were

recorded both by their context number and their sector in the gully. In the south-east part of the gully (Sectors 3, 4) the external edge sloped shallowly outwards to make the gully about 2m wide at that

point. The primary fill of the gully was a grey clayey

silt (C7). Pieces of iron slag and coarse pottery were found in this layer. A radiocarbon date (UB-6996) was obtained from a hazelnut shell fragment from C7 which when calibrated has a 95.4% probability of

being within the range AD 720 to 940 (Table 1). C7 was beneath a charcoal-rich clayey silt (C4), which contained coarse pottery, iron slag and some small

fragments of burnt bone, especially on the east and

south sectors of the gully. Two radiocarbon dates

121



5

were obtained from this context. Grain provided a radiocarbon date (UB-6994) which when calibrated has a 95.4% probability of being within the range AD 720 to 940 (Table 1). Hazelnut shell provided a radiocarbon date (UB-6995) which when calibrated has a 95.4% probability of being within the range AD 660 to 780 (Table 1).

In the north and north-east sectors of the circular

gully (C3) there was a thin deposit of charcoal (C8) above the charcoal-rich clayey silt (C4). In the west and north-west of the gully there was a deposit of iron pan (C27) above the charcoal-rich layer (C4). The south and south-east of the gully was packed with medium-sized stones and pebbles (C26), which were above and within the charcoal-rich clayey silt

(C4). The stone packing (C26) spread out of the

gully to the south and was cut by the later ditch (C9). There were a number of concentrations of pottery within this packing. These appear to have been

substantial fragments of pots which were crushed between the stones.

A second gully (C13) ran northwards from the north-west (Sector 7) of the main gully (C3), and curved to the east. It probably once joined with the main

gully (C3) but the area between the two was interrupted by a field drain (C19). This second gully (C13) was filled by the orange grey, charcoal-flecked,

sandy loam (C14). A third gully (C24) ran north

westwards from the north-east (Sector

1) of the main gully (C3). It was filled

by a light grey loamy clay (C25), which was the same as the basal fill (C7) found in the main gully (C3). It is certain that the gullies (C3) and (C13) were stratigraphically contemporary and likely that all three gullies were

contemporary.

Running east to west across the

southern end of the site was a ditch

(C9), 1.5m wide and 1.0m deep. Its basal fill was a layer of grey clayey silt with charcoal flecks (CIO) which was below a similar but more mottled layer of grey-brown silty clay (C28). This mottled layer may have been ditch-fill mixed with wash from a (now absent) bank. Grain from the basal fill provided a radiocarbon date (UB-6997) which when calibrated has a 95.4% chance of

being within the range AD 1280 to 1400

(Table 1). To the north-east of the excavated

area there was an elongated pit (C5), 1.40m by 0.80m and 0.35m deep. It was filled by a dark grey charcoal-rich silty loam (C 6).

The archaeological features described above were

cut by two field drains. The first of these (C30) was some 0.50m wide and ran in a northerly direction across the entire area under investigation. It was

filled by rounded, well-sorted stones averaging 0.10m in diameter. A second drain in the north of the site (C19) was wider, some 1.5m, and 0.60m deep and ran in a westerly direction, joining at right angles with drain (C30). It was filled with stones at the base and earth at the top. The northerly field drain (C19) cut through, and disturbed, the area where the main

gully (C3) and the second gully (C13) joined. A third drain to the extreme south of the site did not interfere with any of the archaeological features.

122



5

Fig 4 Plan of excavated features.

RADIOCARBON DATING Four radiocarbon dates were obtained, three from the fills of the gully (C3) and one from the fill of the ditch (C9) (Table 1)

The date UB-6995, obtained from material from C4, the upper fill of the gully (C3), is significantly earlier than either the date UB-6994, also obtained from C4 or the stratigraphically earlier UB-6996, obtained from material from C7, the lower fill of the gully (C3). This may indicate that the gully was cleaned out on one or more

occasions. The hazelnut shell from which UB-6995 was obtained could have been residual material from an early clean-out

which got mixed in with the later phase of

silting. This indicates that the gully could have been in use for a significant period of time. An OxCal 3.10 estimate of the span of time between the deposition of the materials from which dates UB-6995 and UB-6994 were obtained suggests that the actual gap between the deposition of these

samples has a 95.4% chance of being within 0 and 180 years and a 68.2% chance of being within 10 and 130 years. This span

Lab No. Context Material Date ,3C Calibrated Range Oxcal 3.10 (Bronk Ramsey 2005); Intcal04 (Reimer et al 2004)

UB-6994 C4 fill of gully C3 Cereal grain 1192 ?29 ^26 AD 720 to 940 UB-6995 C4 fill of gully C3 Hazelnut shell 1287?29 -28 AD 660 to 780 UB-6996 C7 fill of gully C3 Hazelnut shell 1193?29 -27 AD 720 to 940 UB-6997 C10 fill of ditch C9 Cereal Grain 636?28 -24 AD 1280 to 1400

Table la Radiocarbon dates from Terryhoogan.

Atmospheric data from Reimer et al (2004); OxCal v3.10 Bronk Ramsey (2005); cub r:5 sd:12 prob uspfchron]

UB-6994 C4 1192 ?29

UB-6995 C.4 1287?29

UB-6996 C.7 1193?29

U?6997 CIO 636?28 _I_i_I_I_I_I_ _L _L

500CalAD 1500CalAD lOOOCalAD

Calibrated date Table lb Graphical of calibrated ranges of the radiocarbon dates from Terryhoogan.

123



? ? i Fig 5 Sections ?=? ?, -;:_^ 7 .: .:.; z^jo===z^? -? of excavated

X--XJ * o(?D X features

C DE F

G

L

28

^/

R Q

^^^^^^ ^_1 metre

indicates that it is possible that the gully was being used over a period of up to a century or more. If that

was the case, and the radiocarbon evidence neither

proves nor disproves this, then any structure inside

could have been replaced one or more times during the period of use of the gully.

THE POTTERY A total of 506 sherds of pottery were found during the excavation. These sherds almost all belonged to the souterrain ware tradition, and the remainder were

a few fragments of 18th- and 19th-century pottery from the topsoil. Of the sherds found, 427 were body

124



sherds, 51 were rim sherds and 28 were base sherds.

The 51 rim sherds represented approximately 20 vessels. The base sherds represented approximately nine vessels. It is very difficult, especially with coarse hand-made vessels, to match rim and base

sherds successfully if it is not possible to reconstruct a complete profile. Therefore it is impossible to say with certainty if the nine vessels represented by the base sherds are amongst the 20 vessels represented

by the rim sherds. It is, however, likely that many of them are. It can only be said that the 508 sherds from

Terryhoogan are likely to represent between 20 and 29 vessels.

Methodology of the examination of the pottery Little work has been carried out on the Early Christian coarse potting tradition in Ulster. However, an analysis of the vessel forms (Ryan 1973) has defined and provided an outline typology and

chronology of souterrain ware. Some analysis of the

inclusions within the fabric of coarse pottery has been carried out using both microscopic examination

of thin sections of the pottery body (McCorry 1988) and hand lens examination of pottery fabrics (McMullen 2001). The conclusion of these studies has been that the inclusions in the various fabrics have closely matched the background geology of the area, indicating that the pots were made from local

clay and not widely traded. The study of the coarse

pottery tradition of Early Christian Ulster is still in the data acquisition stage and awaits consistent

description and analysis of the full range of body forms and decoration present in the archaeological record.

A methodology focusing on systematic description of the body form and dimensions of individual vessels has been used for the analysis of the coarse pottery assemblage from Terryhoogan. Central to this approach is the breaking down of the

pottery form into elements which can be combined to reconstruct an analysable model of the shape of a

pot. These elements are:

Base diameter

Base/body profile Base form

Rim diameter

Neck/upper body profile Rim form Profile height

The base diameter and rim diameter are measured

with the aid of a rim/base diameter chart. Body shape pro-formas are used to describe the base/body and neck/upper body parts of the pot profile. The forms of the rims and the bases are codified into a

number of basic types against which the pots are

compared. In the small number of cases where it

may be possible to reconstruct a complete profile of a pot the profile height is recorded. Thickness, decoration and colour are also recorded.

Only rims, bases and complete profiles of pots are useable in this form-based classification. Most

body sherds, unless they can be shown toj?in a base and rim or are interesting for another reason, such as

decoration, have not been closely examined.

So as to avoid duplication of data, the rims and bases were grouped together into groups of like sherds so that only one representative rim or base

sherd from each vessel was actually examined, thus

avoiding distortions in the database caused by one

pot being recorded several times because it had broken into more pieces than another.

The information was recorded in a specially designed Microsoft Access database and exported to SPSS 10 for analysis.

A description of the forms of the Terryhoogan pottery Base/body profile forms

Base body/profiles were measurable for nine vessels.

A total of 33% (3 cases) of base/body profiles slanted inwards slightly (up to 5? inwards from vertical) (Fig 6a). A similar number of vessels had

flaring base/body profiles (between 5? and 10? outwards from vertical) (Fig 6b). The next most common form with 22.2% (2 cases) of the observed bases was a flat base with the body flaring outwards

considerably (between 10? and 15? degrees from vertical) (Fig 6c). The least common form, accounting for 11.1 % ( 1 case) of the observed bases, had a flat base and the sides sloping more markedly inwards (between 5? and 10? inwards from vertical) (Fig 6d).

3

2

a b c d

Fig 6 Relative frequencies of base/body profile forms.

125



abed e f g

Fig 7 Relative frequencies of neck/body profile forms.

Neck/body profile forms

Neck/body profiles were measurable for 14 vessels. The most common neck/body profile, accounting for 29.4% of vessels (5 cases) belonged to vessels with

slightly flaring sides and neck (both making an angle of up to 5? outwards from vertical) (Fig 7a). The next most common neck/body profiles each represented 17.6% of vessels (3 cases). These were pots with

slightly flaring sides and an even more flared rim (sides making an angle of up to 5? outwards from vertical; neck making an angle of up to 10?) (Fig 7b), and pots with a widely flaring upper body and neck (both making an angle of 15? to 20? degrees outwards from vertical) (Fig 7c). A total of 11.6% (2 cases) of the observed vessels were slightly globular with a slightly flaring neck (the body making an

angle of up to 5? inwards from vertical and the neck

making an angle of up to 5? outwards) (Fig 7d) or with slightly flaring sides and an in-turning neck (the body making an angle of up to 5? outwards from vertical and the neck making an angle of up to 10? inwards) (Fig 7e). A total of 5.9% (1 case) of the observed neck/body profiles belonged to straight

sided vessels with a slightly inturning neck (body making an angle of 0? from vertical and neck making an angle of 5? inwards from vertical) (Fig li) and vessels with slightly-flaring sides and a widely flaring neck (body making an angle of 5? outwards from vertical and the neck making an angle of 20? outwards from vertical) (Fig 7g).

Base form

Only two of the bases had any form of visible base treatment. One base showed grass marking,

presumably from having been left to dry before firing on grass clippings. A second base bore the impression of basketry. This may have been the result of having been left to dry on a piece of wicker.

Rim form

Only two rim forms were present in the Terryhoogan assemblage; these were the 'rounded' and 'flattened'

forms. Out of the 20 cases with recordable rim forms there was an equal distribution of both forms with 50% (10 cases) rounded and 50% (10 cases) flattened.

Vessel thickness

The vessels at Terryhoogan ranged from 5mm to 12mm in thickness. The mean thickness was 7.9mm and the modal thickness was 8mm.

Rim diameter

The rim diameters of the Terryhoogan pots ranged between 0.14 and 0.24m (Fig 8). There were two modal rim diameters of 0.16 and 0.24m, both with 25% (4 cases) of the measurable rims. The mean value was 0.1975m.

Base diameter

The base diameters of the Terryhoogan pots ranged between 0.10 and 0.20m with a modal value of

Rim diameter Base diameter

Fig 8 Frequency of rim diameters. Fig 9 Frequency of base diameters.

126



Pl 1 Sherd with applied cordon and applied vertical decorative strip.

PI 2 Sherd with wicker impression on exterior surface.

0.10m (Fig 9). The mean base diameter was 0.1375m.

Decoration

The Terryhoogan pottery assemblage displayed a number of types of decoration. Three vessels have

applied cordons below the rim. This is a common feature on souterrain ware (Ryan 1973). The cordons were not very securely attached and probably had a decorative rather than a utilitarian purpose. Two of

the vessels with cordons also have applied vertical decorative strips running up to and joining the cordons (PI 1). One vessel had two lugs applied just below the rim. Lugs are not commonly found on

souterrain ware but are known from Lissue, Co

Antrim (Richard Warner, pers comm). Two body sherds display wicker impressions which may be deliberate decoration (PI 2) or may result from an unfired vessel being left against a wicker panel.

Wicker impressions have not been found previously on souterrain ware. A single vessel displayed a raised section on its rim which appears to be a skeuomorph of the metal mount of a wooden bucket (PI 3). This is also a feature which appears not to have been

found previously on souterrain ware. A single sherd

had a thin groove incised just below the rim. This sherd had a rather different fabric from the other sherds in the assemblage and it is possible that it could be residual from earlier activity in the area.

A suggested typology Because of the small number of individual vessels

represented in the Terryhoogan assemblage it was not possible to suggest meaningful and statistically reliable correlations between different attributes of the data. However it is possible to arrive at an overview based on simple observation of the

assemblage. Of the approximately 20 vessels represented by

the rim sherds from Terryhoogan about 12 belonged to vessels which, to a greater or lesser extent, had

flaring upper bodies and necks. Five vessels had

inturning upper bodies and necks. The size of the vessels was fairly evenly spread with rim diameters between 0.16m and 0.24m and base diameters

between 0.10m and 0.20m. The flaring and inturning vessels seem to be equally spread between the larger and smaller rim diameters. The two rim forms, rounded and flattened, seem to be spread between

the flaring and inturning vessels of different diameters. Decoration and cordons were observed

only on vessels with flaring sides.

PI 3 Rim sherd with protrusion from the rim; possibly a skeuomorph of a bucket handle mount.

127



From this it is possible to propose two basic vessel types from the Terryhoogan pottery assemblage:

a. Vessels with flaring sides, rim diameters

evenly spread between 0.16 and 0.24m and both rounded and flattened rims. Many of these vessels are decorated, some with

cordons and others with impressed dots on the rim.

b. Vessels with vertical or slightly globular bodies and in-turning necks, rim diameters

evenly spread between 0.16 and 0.24m and both rounded and flattened rims. Only one of these vessels displays any elaboration or

decoration, in the form of one lug attached to the side of a vessel just below the rim.

To date only one other souterrain ware assemblage, from Drumadoon, Co Antrim, has been studied using this form-based analytical system (McSparron &

Williams forthcoming). This assemblage was much

larger, consisting of approximately 78 vessels. The vessels from Drumadoon had a larger range of sizes with rim diameters from 0.12 to 0.36m. The vessels from Drumadoon were subdivided into four classes.

A. Small (rim diameter up to 16cm) in-turning bowls with rounded rims

. Small (rim diameter up to 16cm) out-turning bowls with flared rims

C. Larger (rim diameters from 20 to 27cms), vessels, some with inverted rims

D. Very large (rim diameters above 27cms) straight sided vessels with slightly flaring necks and both inverted and pointed rims.

The radiocarbon dates for the two phases in which most of the souterrain ware from Drumadoon was

found (UB-6414 1185+35 BP: 720 to 950 Cal AD; UB6417 1186? 30 BP: 720 to 950 Cal AD; and UB 6418 1199+ 30 BP: 710 to 940 Cal AD) are very similar to two of the three dates for the Terryhoogan assemblage; as discussed above the outlying Terryhoogan date may possibly be the result of residual deposition.

Many other assemblages of souterrain ware have

been found at Early Christian sites all over east Ulster and to comment on all would be outside the

scope of this paper. However, one large assemblage from a well-stratified sequence at Tullylish, Co Down (Ivens 1987), only about 5km from

Terryhoogan, deserves further comment. Although the author does not give the total number of vessels recovered, he appears to describe 55 separate vessels

spread over ten fabric styles. The pottery from the outer ditch was particularly well stratified and dated. The earliest fill of the ditch to contain souterrain

ware pottery (6.19) was radiocarbon dated (UB-2672 1210? 60: 680 to 980 Cal AD). This layer contained

both plain souterrain ware and vessels with cordons

and decoration on the rim. Decorated 'pie-crust' cordoned souterrain ware was found in the two

uppermost layers of the ditch (6.12 and 6.11) along with medieval Ulster coarse pottery (Everted-rim ware). The uppermost of these two layers was dated

by numismatic evidence to the early 14th century (Ivens 1987,119)

There are notable differences between the

Terryhoogan and Drumadoon assemblages. Some of

the difference may be because the Drumadoon

assemblage is much larger and consequently

displays a wider range of sizes and body forms. However, the absence of cordons or other decorative

motifs on the vessels from the 8th- to 10th-century phases at Drumadoon signifies that it belongs to a different strand within the souterrain ware potting tradition. By contrast there are similarities in decorative motifs between the material from the 7th to 10th-century layer at Tullylish and the

Terryhoogan pottery, both displaying cordons and rim decoration. However, as no detailed form

analysis has yet been carried out on the Tullylish material it is difficult to say exactly how the shapes and sizes of the vessels relate to each other. The

similarity of the dates between Terryhoogan,

Tullylish and Drumadoon gives us a snapshot of souterrain ware styles across Ulster at something

fairly close to an instant in time, albeit calibrated to sometime between the very end of the 7th century and the end of the 10th century. At this point it is clear that vessels at Tullylish and Terryhoogan had cordons and rim decoration and vessels at

Drumadoon in north Co Antrim were plain. It has been suggested by Ryan that the earliest

souterrain ware was plain with no decoration or

other elaboration. Decoration appeared later, in the

form of cordons in south Antrim and Down and rim

top decoration in north Antrim. A final stage of vessels with decorated cordons then emerged in the final years of the use of souterrain ware (1973,627 8). Ryan specifically discussed developments in souterrain ware pottery in Antrim and Down because

of the availability at that time of excavated sites with

good stratification in these counties. It seems likely, given the evidence from Terryhoogan, that Ryan's comments regarding the development of souterrain ware in south Antrim and Down are relevant to

Armagh also.

Most of Ryan's conclusions regarding souterrain

ware are still valid today. It may be possible to add to Ryan's scheme and to postulate two general souterrain ware areas in Ulster, a southern and a

northern. The southern area encompasses south

128



Antrim, Down and Armagh, where plain cordons occur, and the northern area encompasses north

Antrim and possibly east Derry, where it appears that decoration with plain cordons is not found (Ryan 1973, 638), although more well-stratified excavations with good dating evidence are needed

to be sure of this. This scheme would see both types of souterrain

ware as initially plain, with cordons and rim decoration, as seen at Terryhoogan, appearing in

south Ulster between the 8th and 10th century. North Ulster souterrain ware would stay free of any elaboration or decoration for a longer period of time, as yet unspecified, before the rim-top decoration discussed by Ryan appears. Later, as suggested by Ryan, decorated cordoned vessels became common

in both areas. Evidence from Drumadoon

(McSparron 2004, 15; McSparrow & Williams

forthcoming) and Tullylish (Ivens 1987, 93, 119) seems to suggest that decorated cordoned souterrain

ware may not appear until the mid-13th century.

METAL ARTEFACTS Philip Macdonald*

Only two fragments of ironwork were recovered

during the excavations. Following Ottaway (1992, 493) both are described as strips, although they probably represent the tip of a nail and either part of a structural fitting or the handle of a small tool.

Curved and tapering strip fragment C4; Sector/Small Find No 3 (derived from Soil

Sample No 2). Curved, rectangular-sectioned strip fragment which tapers to a point. Surviving length 20mm. Possibly the broken tip of a nail.

Elongated looped strip fragment C4; Sector 3; Bag No 44.

Rectangular-sectioned strip fragment which has been turned and closed to form a tapering, elongated loop (internal dimensions 53mm 7mm). The loop is formed by a long, apparently hammer-welded, scarf

joint. The strip is broken immediately below the end of the join. Surviving length 91mm, width at break 7mm, thickness at break 4.5mm. Probably part of a structural fitting or the handle of a small tool.

Burnt bone fragments A number of fragments of burnt bone were found in C4, the fill of the circular gully (C3). Emily Murray of the School of Geography, Archaeology and

Palaeoecology, examined the burnt bone fragments from the excavation but deemed them to be too small to allow identification of the species or other

analysis (Emily Murray, pers comm).

PLANT REMAINS (Table 2) Gill Plunkett**

A total of 151.5kg of soil samples were taken during the excavations at Terryhoogan. Sieving and flotation was carried out by John Davidson of the School of Geography, Archaeology and

Palaeoecology, Queen's University, Belfast. Twelve

samples from seven contexts were found to contain

charred archaeobotanical material. Analysis of these

* Centre for Archaeological Fieldwork, Queen's University, Belfast

** School of Geography, Archaeology and

Palaeoecology, Queen's University, Belfast

Feature Gully C3 Ditch Gully Gully Field drain C9 Cil C13 C19

Context 4 4,7 7 26 10 12 14 20 Samples 1,9,6 11 13 43 21,23 31 33 36,39

Sample weight (kg) 11.49 3.08 3.05 7.2 8.93 4.67 3.32 5.85

CEREALS Cerealia indet. Hordeum spp. Avena spp.

TOTAL CEREALS

SEEDS Corylus avellana nut fragments

Prunus cf. spinosa Unidentifiable seeds

Table 2 Plant remains.

129



remains was funded by the Environment and

Heritage Service: Built Heritage. All archaeobotanical remains were examined using a

Nikon binocular microscope under a magnification ofxlOto x60.

A small archaeobotanical assemblage was

recovered, consisting mainly of charred Avena and Hordeum grains and occasional fragments of

Corylus avellana. Avena and Hordeum are the

predominant cereal types found on both secular and ecclesiastical sites of the Early Christian and medieval periods in Ireland (eg Monk 1986; 1991; Plunkett et al 2005). The assemblage from

Terryhoogan is too meagre to draw any inferences

about crop husbandry at the site, and would appear to represent incidental inclusions in the fill deposits, rather than any specific crop-related activity.

DISCUSSION The main features excavated at Terryhoogan were the circular gully (C3) and its two curving offshoots

(C13, C24). The circular gully (C3) contained some coarse pottery sherds of the souterrain ware tradition. It showed no evidence of having contained structural timbers or wicker and therefore is unlikely to have been a wall foundation. It could, however, have been a drain or drip channel around an Early Christian round house. The other two curving gullies (C13, C24) may have been the drain around an annexe to the main building. Drainage gullies have been found around several Early Christian houses, such as those investigated at Crossnacreevy rath, Co

Down (Harper 1974, 34), Dromore ring-work (a

probable rath), Co Antrim (Collins 1968,61-62), and

Killyliss rath, Co Tyrone (Ivens 1984, 22). The

drainage gully at Crossnacreevy ran for

approximately 5m around the south of a sub-circular

structure defined by a series of post-holes. The gully was up to lm wide and 0.10m deep (Harper 1974, 34). The drainage gully curving around the structure at Dromore ring-work was approximately 10.6m

long, 0.25m wide and up to 0.27m deep (Collins 1968, 61-2). The excavator was of the opinion that the gully, and much of the interior of the site, had been truncated by the removal of approximately 0.30m of soil and that the gully had originally been much wider and deeper. At Killyliss a curving drainage gully approximately 6m in length, 0.40m wide and 0.10m deep partly surrounded an area with a diameter of approximately 5m (Ivens 1984, 22). These were all narrower and shallower than the main

circular gully (C3) at Terryhoogan and none of them made a complete circle. However, these examples are all from rath excavations where the rath ditch would effectively have formed a large sump around the entire site, thereby helping to keep it free from

water. The site at Terryhoogan appears not to have

been enclosed and may have been damp in antiquity. It may be of significance that there were a number of field drains, of modern date, running through the excavated area, presumably to assist with drainage of water from this section of the field.

A layer of large stones (C26) filled the south and south-east of the circular gully (C3), particularly at the point where the gully widened. If the circular

gully (C3) was indeed part of a drip channel or drain around a round house, these stones could have

formed a causeway, through which water could still drain, but which would provide access to the round house inside. The presence of large amounts of

pottery in this area could be explained by rubbish

being thrown or swept out through the entrance of the round house, which was probably located on the south-east of the building

- a typical entrance

position in Early Christian round houses (Lynn 1994, 92).

If the circular gully (C3) was the remains of a

drainage gully around a round house then the house itself must have been slightly smaller than the internal diameter of the gully, which averaged approximately 8.0m. A round house a little less than 8.0m in diameter would have been within the typical range of Early Christian round houses, which are between 4m and 10m in diameter (Lynn 1994, 90). Although the excavation was not extended to follow the curving gullies (C13, C24), it seems likely that

they enclosed an area of approximately 5m by 6m. This would suggest that an annexe of less than this size could have been attached to the main part of the structure. This compares with the size, 4m to 5m,

given for a 'backhouse' in the 7th-century legal tract Cr?h Gablach (Lynn 1994, 90). The absence of structural remains within the area enclosed by the circular gully (C3) may be explained by the deep ploughing known to have taken place at the site in the recent past.

A circular house at Terryhoogan inside a circular

gully (C3) would be compatible with the calibrated

ranges of the radiocarbon dates obtained from the

gully fills. The radiocarbon evidence suggests that the curving gully was in use sometime between the

mid-7th and the mid-10th century. Circular houses were the common house type in the earlier part of the Early Christian period but were replaced by rectangular houses between AD 800 and 1000 (Lynn 1994,85). The radiocarbon dates are also compatible

with the pottery from the site, which is the slightly later 'cordoned' variant of souterrain ware, believed

to have been introduced at an unspecified date in the 'later' Early Christian period (Ryan 1973, 628).

The large ditch (C9) which ran roughly east/west across the south of the site, cutting the circular gully

130



(C3), can be shown by radiocarbon dating to be of the later 13th or 14th century. This demonstrates that it cannot have had been part of an enclosure or

boundary which was directly associated with the circular gully (C3) and any presumed structure inside it. The exact function of the ditch is uncertain. It is located, like the rest of the site, on land which is gently dropping to the south. It is not certain to

which side of the ditch the bank lay. The fill of the ditch suggests that infilling took place largely from the southern, down-slope side. Also the current

townland boundary is located approximately 200m south of the ditch and is defined by a river, the most common boundary marker in early Ireland (Kelly 1997,409-10), which suggests that the ditch is not an earlier manifestation of this boundary. The dimensions of the ditch are consistent with that of the clas or Early Christian ditched field boundary described in the farming law text Bretha Comaithchesa (Kelly 1997,371-2), which was three feet deep, three feet wide at the top and one foot wide at the base. Although this feature is medieval it seems likely that the dimensions of a boundary ditch, intended to demarcate lando wnership, and

presumably curtail the movement of livestock, may not have changed significantly in the intervening centuries. Ditches are mentioned as property boundaries in a number of medieval documents. The

register of Archbishop Alen (McNeill 1950) describes a number of properties where the boundaries are delineated by field ditches as well as

roads, streams and trees. Similar references can be

found in the Ormond deeds (Curtis 1932).

ACKNOWLEDGEMENTS Assistance during the course of the excavation and the

preparation of the report was provided by Dr Colm Donnelly, Keith Adams, Nick Beer, Naomi Carver, Eamon Donaghy, Sarah Gormley, Clare Mc Granaghan, Ronan McHugh, Peter

Moore and Brian Sloan of the Centre for Archaeological Fieldwork, School of Archaeology and Palaeoecology, Queen's University, Belfast. Assistance with post-excavation and publication was provided by Dr Colm Donnelly, Ruth

Logue and Dr Philip Macdonald, of the Centre for

Archaeological Fieldwork, and by John Davidson, Dr Emily

Murray and Dr Gill Plunkett of the School of Geography,

Archaeology and Palaeoecology, Queen's University, Belfast. Helpful comments regarding the nature of the

archaeological features uncovered were made by Dr Chris

Lynn, and Dr Thorn Kerr assisted greatly with helpful comments on the nature of medieval land division. Richard

Warner, formerly of the Ulster Museum, kindly provided assistance in the identification of the pottery.

REFERENCES Archaeological Excavation Unit 2001 Excavation Standards

Manual. Belfast.

Collins, A E 1968 'Excavations at Dromore ring-work', UJA3? (1968), 59-66.

Curtis, E 1932 The Calendar of Ormond Deeds. Dublin.

Day, A & McWilliams, (ed) 1990 Ordnance Survey Memoirs of Ireland, Vol 1, Parishes of County Armagh 1835-38. Belfast.

Edwards, 1990 The Archaeology of Early Medieval Ireland. London.

Harper, A E 1974 'The excavation of a rath in

Crossnacreevy townland, County Down', UJA 36 & 37

(1973-74), 32-41.

Ivens, R 1984 'Killyliss rath, County Tyrone', UJA 47

(1984), 9-35.

Ivens, R 1987 'The Early Christian monastic enclosure at

Tullylish, Co. Down', UJA 50 (1987), 55-121.

Lynn, C J 1994 'Houses in rural Ireland, A.D. 500-1000', UJA 57 (1994), 81-94.

McNeill, C (ed) 1950 Calendar of Archbishop Alen's

Register, c. 1172-1534. Dublin.

McSparron, C & Williams, forthcoming 'The excavation of an Early Christian rath with later medieval occupation at Drumadoon, Co Antrim', PRIA, Section C.

Monk, M A 1986 'Evidence from macroscopic plant remains for crop husbandry in prehistoric and Early Historic Ireland: a review', Journal of Irish Archaeology 3 (1985/6), 31-36.

Monk, M A 1991 'The archaeobotanical evidence for field

crop plants in early historic Ireland', in Renfrew, J M

(ed), New Light on Early Farming: Recent

Developments in Palaeobotany, 315-328. Edinburgh. Muhr, 2001 'Territories, people and place-names in Co.

Armagh', in Hughes, A J & Nolan, W (ed), Armagh History & Society, Interdisciplinary Essays on the

History of an Irish County, 295-331. Dublin.

Ottaway, 1992 Anglo-Scandinavian Ironwork from 16-22

Coppergate (The Archaeology of York, vol 17; The Small Finds r5).York.

Plunkett, G, Bentley, & Collins, 2005 'The archaeobotanical remains', in White Marshall, J &

Walsh, C, Illaunloughan Island: An Early Medieval

Monastery in County Kerry (Appendix 4): supporting archive appendices, available as digital file only; http ://www.wordwellbooks .comiftp I wordwell

/illaunloughan. Bray. Reeves, W 1847 Ecclesiastical Antiquities of Down, Connor,

and Dromore. Dublin.

Reimer, J, Baillie, M G L, Bard, E, Bayliss, A, Beck, J W, Bertrand, C, Blackwell, G, Buck, C E, Burr, G, Cutler,

, Damon, E, Edwards, R L, Fairbanks, R G, Friedrich, M, Guilderson, , Hughen, A, Kromer, ,

McCormac, F G, Manning, S, Bronk Ramsey, C, Reimer, R W, Remmele, S, Southon, J R, Stuiver, M, Talamo, S, Taylor, F, W van der Plichi, J &

Weyhenmeyer, C E 2004, 'IntCal04 terrestrial radiocarbon age calibration, 0-26 cal kyr BP', Radiocarbon 46 (2004),1029-58.

Ryan, M 1973 'Native pottery in early historic Ireland', PRIA13C (1973), 619-45.

School of Archaeology and Palaeoecology, Queen's

University, Belfast 2003 Excavation Manual. Belfast.

This paper is published with the aid of a grant from the Northern Ireland Environment Agency.

131



Excavations at Terryhoogan , Co Armagh: Early Medieval house

Documents