An pe
n arch eriphe
haeo ery o
M a
G r o n i n g
M e
Achl zoolo of a D
a s t e r t h e P r e ‐ & P g e n I n s t
A . G . S 1 e n t o r : W 1 4
lum 2 ogica Dutch
e s i s A r c P r o t o h i s
i t u t e o f J . H u l l e 1 4 5 8 5 0 7 W i e t s k e
‐ 1 2 ‐ 2 0 1
2009 al ana h terp
h a e o l o g s t o r y f A r c h a e
g i e
P r u m m e 0
alysis p sett
y
o l o g y
l
1
s of t tleme
| P a g e
the
ent
2 | P a g e
Summary
1 Preface 4
1.1 introduction 4
1.2 Landscape development and occupation history 5
1.3 Research methods 7
2 Results 9
2.1 Number of remains 9
2.2 Wither heights 12
2.3 Natural environment and site surroundings 15
3 Diet 17
3.1 Livestock composition 17
3.2 Livestock consumption 18
3.3 Fish 21
3.4 Poultry 22
3.5 Molluscs 22
4 Age at Death and secondary products 24
4.1 Cattle 24
4.2 Sheep 29
4.3 Other domesticated mammals 31
5 Anomalies, pathologies and tools 33
5.1 Scurred cattle 33
5.2 Achondroplastic dog 36
5.3 Other anomalies and pathologies 37
5.4 Tools 38
3 | P a g e
6 Special depositions and features 40
6.1 Special depositions 40
6.2.1 Animal remains in features: introduction 42
6.2.2 Animal remains in features: rainwater catchment depressions 43
6.2.3 Animal remains in features: pits and wells 44
6.2.4 Special features: ditches and trenches 45
7 Regional Setting 46
7.1 Introduction 46
7.2 livestock 46
7.3 Non livestock species and settlement status 51
8 Conclusion 54
Literature 58
Appendix A: Glossary of species 61
Appendix A: Location of the terp and elevation maps of the region 63
Appendix B: Meat Weight per period 65
Appendix C: Marked, pathological and worked remains 66
Appendix D: Distribution of skeletal elements of livestock species 68
Appendix E: Excavated skeletal elements of the partial cattle skeleton 71
Appendix F: Excavated skeletal elements of the partial sheep skeleton 72
4 | P a g e
1 Preface
1.1 Introduction
From 2 June until 25 June 2009 the excavation of the artificial dwelling mound site Achlum took place. These artificial dwelling mounds are commonly named terps in the province Friesland and wierden in the province of Groningen. The excavation was commissioned by the Groningen Institute of Archaeology (GIA) and was led by Dr. J.A.W. Nicolay. The excavation is part of a larger research project in collaboration with the province of Friesland focused on the occupation history of the Frisian coastal area in relation to the development of the landscape. The investigators wanted to examine the way in which the landscape around the settlement was used. The gully that flowed near the settlement is of interest because it was connected to the Marne and therefore had a direct connection to the sea (Appendix B). In what way this gully was used by the inhabitants of the terp is of particular interest. The conservation condition of the terp and the influence of erosion will also be studied (Nicolay & Blok 2009, unpublished).
In this thesis the zoological material was analysed to obtain answers about the relationship between man and animal on the terp. The role of the various animal species for the food supply of the site will be examined as well as the use of other animal products. The evidence for the
production of tools using animal remains will also be studied. The animal remains will be analysed to extract information about the use and variation of the landscape. Because the site was inhabited from the Middle Iron Age to the Late Middle Ages it is possible to examine changes in the use of the landscape and the composition of the livestock over time. Another question that will be answered by analysing the animal remains is the presence or absence of ritual or other special animal deposits.
Finally the information collected by analysing the archaeozoological remains of Achlum will be compared to other terp settlements in the region. The various layers and features of Achlum have been dated using potsherds. This led to the breakdown of the occupation period of the site into a number of phases which can be found in table 1.
Table 1: Habitation periods of Achlum
Phase Period Date
1a Middle Iron Age 400‐200 BC
1b Late Iron Age 200 BC‐ AD 50
2 Roman period AD 50‐250/300
3 Migration period AD 400‐500/550
4 Merovingian period AD 500‐700
5a Carolingian period AD 700‐900
5b Carolingian/Ottonian period AD 800‐1000
5c Ottonian period AD 900‐1100
6 Late Middle Ages A AD 1000‐1200
7 Late Middle Ages B AD 1100‐1300/1400
1.2 Land A Friesland The loca clay‐on‐p formatio age. The resulted North‐w because Boorne t than 5 m layer and
Grey spots
T was that area cha In the co salt‐mar Between this deve and the
dscape devel Achlum is lie d between th
tion of the s peat landsca on of peat du e increase of in a rise of t western part o
the river Bo tidal basin. T meters below d the apex o
s indicate the ter
The sea leve t the Boorne anged into a ourse of the rshes toward n the 200 BC elopment is t
tidal system
lopment and es in an area
he former es settlement te ape formed d uring the Hol the tempera the ground w of Friesland oorne drained The apex of t w NAP. This w
f the Pleisto
rps that are prese
l rise decrea e tidal basin g
salt‐marsh in Iron Age and ds the north.
C and 500 BC that the cree s of the Mar
d occupation named Wes stuaries of th erp of Achlum during the Ho
locene was t ature resulte water level a
that was ove d into this ar he Pleistoce was caused b cene deposit
Figure 1: West
ent in the Early M terp of Achlu
sed in the co gradually fille n which a cre d Roman per
A successio sedimentati eks in the hin rne and Midd
n history tergo that is he Middelzee
m is indicate olocene in th the result of ed in the flow
nd peat was erflowed by rea. Achlum ene sand in th by a tidal gull
ts during the
tergo in the Ear
Middle Ages. Black um is indicated w
ourse of the ed up with se eek system w riod the sedim on of salt‐ma
ion increases nterland are delzee replac
s located in t e and Marne ed by a red do
he area betw the sea leve w of melt wa formed in a sea is also ca is located o he subsoil of ly that erode e Middle‐Hol
rly Middle Ages
k dots indicate th with a red dot.
Holocene. T ediment. At was formed ( mentation re rsh ridges w s the height silted‐up. Th ce the system
he north‐we e and the Wa ot. The site i ween Harling l rise after th ter to the se n area paral alled the ‘Bo n the south‐
f the excavat ed the previo
locene.
s
he location of cur
his result of the end of th (the pre‐Rom esulted in th
ere formed d of the coastl he drainage d m of creeks.
5 estern part o addenzee (fig
is located wi gen and Snee he Weichseli ea. The sea le
lel to the coa oorne tidal ba
‐western flan tion area lies ously present
rrent villages and
this develop he Bronze Ag man Marne s
e expansion during this p line. The res direction cha
| P a g e f
gure 1).
thin the ek. The
ien ice evel rise
ast. The asin’
nk of the s more
t peat
towns, the
pment ge the system).
of the period.
ult of anges
6 | P a g e From 700 BC salt‐marsh ridges were present in the southernmost part of Westergo that were suitable for habitation. The presence of salt‐marsh ridges is one of the most characteristic
morphological features in Westergo. The location of some of the salt‐marsh ridges is given in
Appendix B. The salt‐marsh ridges of the Westergo area run more or less parallel to the present‐day coastline of the Waddenzee. The ridges are generally around 200‐500 meter wide and around 0.5 – 1.5 meter elevated above the surrounding surface if terp layers are not taken into account. The earliest terps in the Westergo area were built on the salt‐marsh ridges. This led to the development of linear settlement patterns (terp rows). Achlum is part of a number of artificial dwelling mounds located on the salt‐marsh ridges. Some of the salt‐marsh ridges are still present and are visible on elevation maps (Appendix B). Other salt‐marsh ridges are no longer visible because they were disturbed by agricultural activities or other causes. The first areas to be occupied were the salt marshes in the southernmost part of Westergo. The salt‐marsh ridges which were formed north of this area were colonised during the Late Iron Age and the Roman period (Vos 1999). In the 4th century AD large parts of Westergo are deserted caused by the movement of populations catalysed by changes within the Roman Empire. After the 4th century AD Westergo is assumed to be inhabited continuously until the present day (Bazelmans 2000).
During the Early Middle Ages the Marne tidal system was reactivated causing the gullies and creeks in the area to expand. This development was the result of peat reclamation and the resulting consolidation of the soil. The influence of the sea on the hinterland increased through the Marne and Middelzee tidal systems. The increase of the storm water storage area resulted in the increase of the water transportation gullies in the Marne area. Along the gullies relatively high salt‐marsh ridges were formed on top of the older salt‐marsh deposits. The resulting landscape is named the medieval Marne system. Between the 700 BC and 1000 BC the consolidation of the soil resulted in the erosion of deep and broad tidal gullies that reached far into the hinterland. After 1000 BC the sedimentation of the hinterland increased the soil consolidation and the area starts to silt up again (Vos 1999). After AD 1000 dikes were constructed in the area and the sedimentation halted almost completely. Only during floods some clay was deposited in the area (Nicolay & Blok 2009, unpublished).
7 | P a g e 1.3 Research methods
The terp was partially destroyed around 1860 during a commercial excavation. Terp soil was in demand because farmers used terp soil to fertilize their land. The archaeological excavation of 2009 was focused on examining the build‐up of the terp by excavating the scarp face and the
surrounding area south of the terp (Appendix B). Adjoining the scarp face a 5 m wide trench was dug to examine the outer area of the terp. Two other trenches were dug at a right angle to the first trench. All trenches were deepened until the natural surface was reached. The profile and trenches were deepened mechanically using an excavator and were trimmed manually. The fillings of pits, ditches and wells were sieved using 5 mm and 2 mm mesh gauges. A total of 51 sieve samples were collected. Sample sizes varied between 10, 12, 24, 36 and 72 litres depending on the size of the feature. Some features were completely collected like the habitation layers of the sunken hut that will be discussed in paragraph 6.2. Other animal remains were collected by hand.
The comparative collection of the Groningen Institute of Archaeology (GIA) was used to identify the species of the remains. Only remains that could be dated accurately were used in the analysis of the animal remains excavated at Achlum. Only a few remains of less frequent species and a few abnormal and worked remains that could not be dated were studied. The bones have been measured using the method that was developed by von den Driesch (von den Driesch 1976) based on the work of Karl‐Heinz Habermehl. A digital vernier calliper was used to minimize measuring errors. A digital scale with a resolution of 0.1 g was used to measure the weight of most of the bones. To measure small fragments like those of the 2 mm and 5 mm fractions a digital precision scale was used with a resolution of 0.01 g.
The animal remains from the various habitation phases were analysed separately to examine the use of animal products over time. Finds that could not be attributed to a specific period have been left out. Phase 5(+2) was excluded because the finds could belong to both phase 5 and phase 2.
Phase 5(?) was also omitted; remains attributed to this phase could belong to the Carolingian, Carolingian/Ottonian or Ottonian period. Partial skeletons were analysed separately to prevent bias.
Bones that most likely belong to a specific period were added to that phase. Phase 2 is combined with phase 2(?) and the finds dated to phase 6, 6a, 6b as well as 6(?) were joined into one phase.
Phase 6‐7 was added to phase 7 because they overlap in time. Phase 3 includes the finds that can be attributed to the Migration Period with some certainty. No animal remains were found during the excavation that can be attributed to this phase although some pottery shards were dated to this phase.
There are a few different methods in use to estimate the wither heights of cattle. The
multiplication factors for metapodia of Von den Driesch & Boessneck were used here to calculate the cattle wither heights (Von den Driesch & Boessneck 1974). Differentiation between the bones of cows, bulls and oxen was not possible at Achlum because there were not many complete bones present. For this reason wither heights for bulls and cows were calculated separately. An average value was also included and is often used to estimate the wither heights of oxen. Wither heights for the other skeletal elements have been calculated using the multiplication factors of Matolcsi (quoted in von den Driesch and Boessneck 1974, p. 336). The multiplication factors of M. Teichert (quoted in von den Driesch and Boessneck 1974, p. 339) were used to calculate the wither heights of sheep. The total lengths of
metapodia of dogs can be used to calculate the wither height using the regression formulas of Clark
8 | P a g e (Clark 1995). The long bones of dogs can be used to calculate the wither height using Harcourt’s
regression formulas (Harcourt 1974). The regression formulas of E. May; M. Teichert and K.
Hanneman were used to calculate the wither heights of pigs (quoted in von den Driesch and Boessneck 1974).
Specimen count is used to quantify and reconstruct the composition of the livestock kept at archaeological sites. This method has been reviewed extensively which has resulted in a few different abbreviations and terms for the same technique. Examples are NR for number of remains (Morales et al. 1994), TNF for total number of fragments (Gilbert and Steinfeld 1977) and NISP for number of
identified specimens (Payne 1975). From this point onwards NR will be used to denote this method.
Partial skeletons will not be included in the quantification of the number of identified specimens and will be analysed separately. Skeletal elements that could be attributed to the same animal have been counted as one. The NR will be shown separately for the material that was collected by hand and for the 2 mm and 5 mm sieve fractions. The sieve fraction remains were collected by sampling specific features while almost all of the hand‐collected remains from the excavated area were collected.
Joining the remains from these different collection strategies would result in misinterpretation of the data. Merging the different groups would result in underestimating the importance of fish and bird remains. The high degree of fragmentation of the sieve fraction remains would also influence the results if merged with the hand‐collected material.
The bone‐weight method as described by Reed will be used to approximate the importance of the different mammal species to the diet of the inhabitants of Achlum (Reed 1963). This method is developed for mammal species and it is therefore not possible to use this method to analyse non‐
mammal species. Reed’s method assumes that the skeletal weight of large mammals is about 7.5% of the total weight of the animal. Secondly the method assumes that the edible meat of an animal is 50% of the total meat weight. This method resulted into the following formula:
Meat Weight= 0.5 x (Bone Weight/0.075)
To determine the age at death of cattle, sheep, pigs and horses, the fusion tables of
Habermehl were used (Habermehl 1975, pp. 57‐132) combined with the comparative collection of foetal and young animal skeletons of the Groningen Institute of Archaeology (GIA). To establish the ages of death of the dog remains found at Achlum the fusion‐ages of H. Reichstein (unpublished),
complemented by the comparative collection of the Groningen Institute of Archaeology. The fusion‐
ages established by Reichstein differ from the fusion‐ages established by Habermehl. Using the same method for all species is preferable but Reichstein’s method includes several skeleton elements that Habermehl doesn’t include in his method. The transition from subadult to adult differs greatly between different dog breeds. Because of this a period of 8 to 18 months has been used as a transition phase instead of using a specific age.
9 | P a g e
2 Results
2.1 Number of Remains
The hand‐collected remains from normal features consist of 1030 remains (table 2). For the most part these belong to livestock species. Because no goat remains were found it is likely that all remains in the sheep/goat group are actually sheep remains. Cattle and sheep were the most important species at Achlum. There are much less pig, horse and dog remains present at Achlum.
Only one fragment of a cat, found in an undated feature, could be identified. Bird remains also were present, but in much lower quantities than the mammal remains. Most of them are goose and mallard remains, but there are also two remains from chicken present in the Carolingian period remains. A number of mollusc shells have also been found among these are mussel and common cockle shells. No fish bones have been found in the hand‐collected material but two fragments of a caudal vertebra of a sperm whale were found in the material from the Late Middle Ages A. A glossary of the English, Latin and Dutch names of the animal species found at Achlum and other terp and wierde settlements can be found in Appendix A.
Table 2: Achlum 2009 ‐ Number of Remains (hand‐collected)
Phase ‐ 1a 1b 2 4 5 5a 5b 5c 6 7
size unknown 4 6 14 5 5 13 5 3 4
sheep/goat/pig size group 1 1 18 5 11 29 12 12 7
horse/cattle size group 1 11 29 5 1 11 16 14 13 9
cattle 2 4 39 87 26 2 12 38 22 47 70
sheep/goat 9 48 73 32 26 22 36 15
sheep 1 5 7 4 1 2 3 2
pig 10 1 2 2 8 1 3 5
dog 2 5 3 1
horse 1 3 1 2 2 4 2
cat 1
Total NR Domestic mammals 8 5 70 218 122 5 78 133 81 124 115
mallard 2 1 2
chicken 2
grey /black geese 1 1
brant goose 1
unidentified bird 1
Total NR Birds 4 0 0 0 2 0 2 0 0 3 0
common cockle * 4 5
peppery furrow shell 9 3
mussel 11 5 2 1 5 1 9
baltic macoma 6
Total NR Molluscs 11 0 20 6 0 0 1 5 3 6 9
sperm whale 1
Total NR Cetacea 1
Total number of remains 23 5 90 228 224 5 81 138 84 128 124
* Not counted, sample from a layer of shells.
10 | P a g e The 5 mm and 2 mm sieve fractions have yielded quite a few remains, most of which could not be identified to species (table 3 and 4). Mostly the same domesticated mammal species as in the hand‐collected material are present with the exception of the cat. Quite a few fish remains were found in the 5 mm sieve fraction. These remains will be analysed further in paragraph 3.3. The 5 mm fraction yielded several chicken remains; one of them is dated to the Roman period. Other chicken remains were found in features attributed to the Carolingian period and in the Late Middle Ages. A few mollusc remains could be identified in features attributed to the Late Iron Age, the Carolingian period and the Late Middle Ages. Most of the mouse remains belong to the vole family. A single bone from a natterjack toad was present in the remains from the Late Middle Ages B.
Table 3: Achlum 2009 ‐ Number of remains (5 mm sieve fraction)
Phase 1b 2 4 5 5a 5b 5c 6 7
size unknown 2 15 1 110 28 14 3 2
sheep/goat/pig size group 2 12 3 3 7 11 14 10
horse/cattle size group 8 10 3 8 3
cattle 1 4 3 4 8
sheep/Goat 10 4 2 28 15 1 19 5
sheep 2
pig 2 1 1
dog 1* 1 1
Total NR Domestic mammals 5 49 8 116 79 49 1 55 20
flatfishes 2 1 7 36 5
european eel 1 2 1
garfish 4
cod 1
carp family 1
unidentified fish 10 7
Total NR Fish 0 2 1 0 19 0 0 50 6
chicken 1 8 2 2
mallard 1
unidentified duck 1
unidentified bird 1 1 2
Total NR Birds 0 2 0 0 9 1 0 3 4
mussel 4 2 2 8
common periwinkle 3
baltic macoma 4 2
common cockle 14 20
Total NR Molluscs 4 0 0 0 5 0 0 20 30
vole family 1
unidentified mouse 1
Total NR wild mammals 0 0 0 0 1 1 0 0 0
natterjack toad 1
Total NR Amfibia 1
Total number of remains 9 53 9 116 113 51 1 128 61
*Partial skeleton foetal dog, counted as one animal.
11 | P a g e The 2 mm sieve fraction (table 4) did not yield any mammal remains of which a species could be determined. One mallard bone could be determined in the material. Some mice and toad remains have been found in the 2mm sieve fraction. Voles were present throughout the habitation period of the terp with the exception of the Iron Age and the Late Middle Ages B. Some of the vole species could be identified to species. Tundra vole was identified in the Roman period, Carolingian period and in the Ottonian period. Two other mice species were not added to the table because they could not be dated precisely enough. The common shrew and the common vole are both present in features that belong to either the Roman period or the Early Middle Ages. The 2 mm sieve fraction shows a great variety of mollusc species, but only the Roman period and Late Middle Ages B features yielded mollusc remains.
Table 4: Achlum 2009 ‐ Number of remains (2 mm sieve fraction)
Phase 1b 2 4 5a 5b 5c 6 7
size unknown 6 2 1 2 9 2
sheep/goat/pig size group 2
indet 2 20
Total NR Domestic mammals 2 6 2 20 1 2 11 2
gadids 1
herring 12 1 2
flatfishes 2 1 10 1 8 3
european eel 2 1 3 1 6 2
salmon/trout 1
garfish 1
three‐spined stickleback 1
unidentified fish 2 4 5 14 1 19 10
Total NR Fish 4 6 7 34 3 0 36 17
mallard 1
unidentified bird 1 1 2
Total NR Birds 0 1 1 0 0 0 1 2
common cockle 2
layer spire shell 8 1
slender amber snail 8
widespread column 4
arctic barrel‐bubble 8
peppery furrow shell
european stream valvata 2
unidentified mollusc 2
Total NR Molluscs 0 22 0 0 0 0 0 13
vole family 1 2 9 8 6
tundra vole 3 1 2
unidentified Mouse 1 2 1 1 2
Total NR wild mammals 0 5 2 12 10 2 7 2
toad family 3
unidentified amphibian 1 7
Total NR amphibian 0 0 0 1 0 0 0 10
Total number of remains 10 41 12 67 14 3 55 46
12 | P a g e 2.2 Wither heights
Only a limited number of cattle remains were complete enough to be used to calculate the wither height (table 5). The limited number of remains of which a wither height could be established make it difficult to analyse changes in the size of the cattle over time. Figure 2 shows the expected value of cattle wither heights over time of Romanised areas. The average wither height of the terp cattle of the Roman period is quite different and is represented in the figure with a blue line. Because the values represented in the figure are averages red lines were added to visualise an estimate range of the wither heights within a period. The wither height of the Late Iron Age metacarpus with find number 275/1 of did not divert much from the expected wither heights of cattle from this period.
The metacarpus with find number 287/2 is quite large for the Late Iron Age. The average size of the cattle of the terp Bullepolder (AD 0‐250) is 107 cm (Halici 2002a).
The wither heights of the cattle of the Roman period are slightly higher than those of the terp settlements of Sneek‐ Stadsrondweg Oost and Wijnaldum. Two wither heights could be established during the analysis of the animal remains of Sneek‐ Stadsrondweg Oost (Halici 2002b). The two metacarpi yielded wither heights of 103 and 113 cm. During the analysis of the animal remains of Wijnaldum three wither heights of cattle could be established (Esser, Zeiler & Prummel, in preparation). Two metacarpii yielded wither heights of 108.5 and 103.9 cm and a metatarsus yielded a wither height of 109.6 cm. A Migration period metatarsus from Wijnaldum yielded a wither height of 111.7 cm and a Merovingian period metatarsus a wither height of 109.2 cm. No cattle bones were found at Achlum in features dated to those periods that were complete enough to establish wither heights.
A metatarsus from the Carolingian period of Achlum yielded a wither height which is comparable to what is expected from that period. Three wither heights from this period based on cattle metatarsi were established during the analysis of the animal remains of Wijnaldum. Another metatarsus from the Ottonian period of Wijnaldum was very large and yielded a wither height of 129.1 cm. The metatarsus with find number 151/5 from a Late Middle Ages feature of Achlum is distinctively smaller than expected. The partial cattle skeleton (f.nr. 144) of the Ottonian Period produced a number of bones suitable for the determination of the wither height. The average wither height is 113.6 cm but this hasn’t accounted for the fact that multiplication factors for different bones were used. The wither height values established using metapodia and the multiplication factors of Von den Driesch & Boessneck are usually slightly higher (Von den Driesch and Boessneck 1974).
The wither height of the animal most likely was around 116 cm considering the fact that the humerus with find number 144/10‐1 yielded a wither height of 116.7 cm.
Period 1b 1b 2 2 5a 7 5c 5c 5c 5c 5c Average
T sheep. T Caroling yielded a the shee of comp height o 63.6, thr
F.nr.
287/2 275/1 276 283/1 189/3+11 151/5 144/34‐2 144/52 144/51 144/34‐1 144/10‐1 e wither heig
F
Table 6 show The data set i
ian/Ottonian a wither heig ep/goat with
lete bones th f 65.8, a met ree metacarp
Figu
T
Skeletal el metacarpu metacarpu metatarsu metatarsu metatarsu metacarpu tibia metatarsu metatarsu femur humerus ght of cattle
F.nr: find numb
ws three skel is too small t n period and ght of 55.5 c er heights of hat could be tatarsus yield pi yielded wi
ure 2: Wither h
Table 5: Achlum
lement Siz us
us s s s us
GL GL GL GL GL GL Feature s
s
GL GL GL GL GL fnr. 144 in c
ber, GL: greates
etal element to say much d the Late Mi m. This is qu f Wijnaldum e used to esta
ded a wither ther heights
height changes m 2009 – withe
ze Size (m L
L L L L L
197.8 181.5 211.06*
211 210.07 171.51 203 – Partia L
L L L L
320.52 215.1 214.9 329.7 282 m
st length, *Not
ts that could about possib iddle Ages A uite small for . The Carolin ablish withe r height of 63 s of 61.7 cm,
of cattle over t
er heights cattl
m) Factor
*
6/6.3 6/6.3 5.3/5.6 5.3/5.6 5.3/5.6 6/6.3 al cattle skele
3.45 5.3/5.6 5.3/5.6 3.23 4.14
precise, wear o
d be used to ble changes . An astragal r a sheep of t ngian period
r heights. A m 3.2 cm, a rad
70.4 cm and
time
e
Cow(cm)
6 6 6
118.7 108.9 111.9 111.8 111.3 102.9 eton 6 6
114.0 113.9
on proximal en
calculate the in wither he us from the this period w of Wijnaldu metacarpus y dius yielded a d 68.5 cm.
13
) Bull(cm) 124.6 114.3 118.2 118.2 117.6 108.1
120.5 120.3
nd
e wither heig eight betwee
Carolingian when compar m yielded a yielded a wit a wither heig
| P a g e Average(c
121.6 111.6 115.0 115.0 114.5 105.5 110.6 117.2 117.1 106.5 116.7 113.6
ghts of n the Period red with
number ther ght of
cm) 6 6 0 0 5 5 6 2 1 5 7
14 | P a g e A calcaneus from the Ottonian period of Achlum, found in the same feature as the partial cattle skeleton, yielded a wither height of 62.2 cm. Finally an astragalus attributed to the Late Middle Ages yielded a wither height of 65.8 cm. Wither heights based on the astragalus and calcaneus are not as precise as those based on the long bones. This is caused by the small size of the greatest length of the lateral half of these bones in when compared with the wither height. With this in mind the calculated wither heights seems to remain fairly constant between the Carolingian/Ottonian and the Late Middle Ages A. During the analysis of Wijnaldum several wither heights of sheep/goat of earlier periods could be established. A metacarpus from the Roman period of Wijnaldum yielded a wither height of 65.3 cm. A tibia yielded a wither height of 66.7 cm and a metatarsus a wither height of 64.9 cm. The Migration period is represented by a single metacarpus that yielded a wither height of 64.1 cm. The Merovingian period yielded a metatarsus that yielded a wither height of 70.6. These remains show that the size of the sheep of Wijnaldum did not change much over time. Whether the size of the sheep of Achlum remained constant in from the Iron Age to the Carolingian/Ottonian period is not clear.
Table 6: Achlum 2009 ‐ Wither heights sheep (cm)
period F.nr. Skeleton elements size size (mm) factor Wither height (cm) 5b
6
62/9 89/2
astragalus astragalus
GLl GLl
24.48 29.9
22.68 22.68
55.5 65.8 Feature 203 – other
5c 144/60 calcaneus GL 54.57 11.4 62.2
F.nr: find number, GL: greatest length, GLl: Greatest length of the lateral half
Only two dog bones could be used to establish the wither heights. The dog metacarpus V found in layers attributed to the Late Iron Age is of a medium size with a wither height of 51.7 cm.
This means that this animal is slightly smaller than a modern German shepherd dog. The tibia of the other dog from the Late Middle Ages A can’t be used to calculate the wither height properly because it is aberrant. If assumed that the regression formulas can be used in this case, the dog would be quite small. A radius found in a Roman period feature of Wijnaldum yielded a wither height of 62.7 cm. A tibia from the Carolingian period of Wijnaldum yielded a wither height of 44.8 cm. The only skeletal element of a pig suitable for establishing a wither height is an astragalus from the Late Middle Ages A. The wither height of this animal is 84.3 cm.
Table 7: Achlum 2009 – wither heights dog (cm)
Period F.nr. Skeletal Element Size Size (mm) Wither height (cm)
1b 355 metacarpus V GL 54.3 51.7
6a 84/9 tibia GL 128.5* 38.5
Fnr: find number, GL: greatest length, *: Achondroplastic dog.
15 | P a g e 2.3 Natural environment and site surroundings
The presence of certain species can yield much information about the environment of a site.
The most useful species are those with a very specific habitat. Livestock is therefore very usable in analysing the natural environment around a site but some livestock species are more adapted to certain environments. The pig for example is more suited to forest environments and sheep are difficult to keep in fresh water environments. Molluscs, birds and fish are more suitable to determine the climate and natural environment of a site in a specific period of habitation. The wild bird remains that could be identified do not provide much information about the natural environment of the site.
The mallard is a (mostly) resident breeding bird which lives near a variety of waters without a specific habitat preference. This species was found in features from the Roman period, the Merovingian period and the Late Middle Ages A. The only bird species that can yield information about a specific habitat is the brant goose. This goose hibernates near the coasts of northwest Europe in the tidal estuaries. It currently hibernates in the Netherlands between September/October and May/June (Animalbase project group). Just one bone of this species was found but not in a feature that could be dated.
Most of the mollusc species live in brackish or saltwater environments. The arctic barrel‐
bubble for example is a burrower that currently occurs near the shore of the North Sea were it lives in the first few cm of the muddy surface and feeds on the layer spire shell. The layer spire shell lives in the silt soils of shallow water in and just below the littoral zone. The layer spire shell can live in water with a lowered salt content up to 10% and can survive for several hours during low tide (Animalbase project group). These species most likely lived in the Marne and were deposited in the vicinity of the terp during floods. Peppery furrow shell, mussel and common cockle also live in the littoral zone of seas. The common cockle can withstand salt concentrations between 15 and 40%
(Animalbase project group). The common cockle is another indication for the deposition of mollusc shells during floods. Most of the remains of this species were found in sedimentation layers of clay within ditches and pools. Although changes in habitat have to be taken into account most of the mollusc species found at Achlum point to a brackish environment.
Four gastropod species were identified: European stream valvata, widespread column, slender amber snail and common periwinkle. The widespread column is a land snail living on dry meadows, sand dunes and open and sunny habitats. This species is currently often found in Britain in sheep‐grazed calcareous grasslands. The slender amber snail is a terrestrial pulmonate gastropod mollusc found near the waterside but unable to survive long underwater. The European stream valvata is a small snail found in clear standing or slowly moving water, up to 80 m deep but
commonly around 3‐10 meter deep. The species lives on muddy or silty substrate and requires high oxygen content. The European stream valvata can withstand water with a salt concentration up to 0.4%. The common periwinkle and baltic macoma are marine gastropods that live in the littoral zone and can withstand very low salt concentrations (Animalbase project group). The gastropods also point to a brackish environment but were most likely living in the gully and the surrounding area that was located near the site.
The fish present at Achlum mostly live in marine or brackish water with the exception of cyprinids which are primarily freshwater species and are only sporadically found in brackish water.
Flatfishes and gadids chiefly live in marine and occasionally brackish waters, but they can survive in
16 | P a g e fresh water as well. Other marine and brackish species are garfish and herring (Froese &, Pauly 2010).
Fish species that migrate from the sea to fresh water to spawn (anadromous) or that migrate from freshwater to the sea (catadromous) are represented as well. The European eel is an example of a catadromous species. This species travels from fresh water to its spawning area in the Sargasso Sea in the Atlantic Ocean near Bermuda and can survive in fresh, brackish or marine waters with
temperatures between 4°C and 20°C. No distinction between the remains of salmon and trout could be made. Both species are anadromous and spend their youth in large rivers after which they migrate to the sea were they reach maturity and return to their spawning grounds. The three‐spined stickleback is another anadromous species that occurs in fresh waters, estuaries and coastal seas. It inhabits shallow vegetated areas usually over mud or sand. It can survive in waters with
temperatures between 4°C and 20°C (Froese &, Pauly 2010). With the exception of the cyprinids, the fish found at Achlum most likely lived in the Marne or in the gully connecting the Marne to the terp. The single cyprinid vertebra that was found in a feature from the Late Middle Ages A could have lived in the freshwater catchment basins on the top of the terp. The species present in each of the periods do not yield information about changes in the natural environment of Achlum.
3.1 Lives
F
Achlum g exceptio the com evenly b and shee separate remains low perc Zanderin was kept day. Goa their sus T cattle gr sheep in again un Although higher th Late Mid remains and the
stock compo Figure 3 show
grouped by on of the poo
plete herd to because of th
ep in most h ely it is likely
could be ide centage of go nk (Zanderink
t to control t ats were also sceptibility to
The Iron Age adually decr ncreases grea ntil it reaches h the percen han before t ddle Ages the
of sheep un percentage o
Cattle Horse Sheep/G Sheep Pig Dog
1
Number of Remains (%)
osition ws the distri period. Ther orly represen
o for exampl he difference abitation pe that most o entified in co oats was pre k 2010) there the sheep. T o kept, accor o infection d e and the Rom reases from t atly. During t s almost the ntage of shee he Meroving ere is a patte til the Caroli of sheep dec
Figure 3: Achlu 1a 100
0 Goat 0 0 0 0 100 2030 4050 6070 8090 100
Achlum Numb
3
bution of the e are multip nted Middle le diseases is es in diet of d riods of Ach f the remain ontrast to a f esent at Achl is a practice his animal w rding to Zand
iseases.
man period s the Late Iron the Carolingi same value ep decreases gian period u ern visible of ingian period creases.
um 2009 ‐ Num
1b 2
75 55, 1,9 1,9 17,3 30,4
1,9 3,2
0 6,3
3,8 3,2
m 2009 ‐ er of Rem
3. Die
e number of ple livestock s
Iron Age. By s reduced. Th different spe
lum. Althoug ns classified a fair number o
um but coul e known from would bring t
derink to pre
show high pe n Age to the
ian/Ottonian as in the Lat s after the Ca until the Late f decreasing d after which
mber of Remains 4 1 24,3
9 0
4 68,2
2 6,5
3 0,9
2 0
Domest mains (h
t
f identified li species pres y keeping mu
he environm cies. There i gh sheep and as sheep/goa of sheep rem d not be ide m the region
he sheep ba edict the outb
ercentages o Carolingian p n period the te Iron Age d arolingian pe e Middle Age numbers of h the percen
s of the domes
5a 5b
23,5 50,7
2 2,7
62,7 34,7 7,8 1,3 3,9 10,7
0 0
icated an hand‐coll
vestock rem ent in all per ultiple specie ment is also ex
s a preferenc d sheep/goat at are actual mains. It is po
ntified to spe that states t ck to their pe breaks of dis
of cattle. The period while percentage o uring the Lat eriod, the pe es. From the cattle remai tage of cattl
ticated animal
5c 6
44,9 49 4,1 4, 44,9 37 4,1 3, 2,0 3, 0,0 3,
nimals:
ected)
17 mains found a
riods with th es the risk of xploited mo ce visible for t are presen ly sheep. No ossible that a
ecies. Accord that a single
ens at the en seases becau
e percentage e the percent of cattle incr te Middle Ag rcentages re Late Iron Ag ins in favour e increases a
ls
6 7
9,5 73,7 ,1 2,1 7,1 15,8
,1 2,1 ,1 5,3 ,1 1,1
| P a g e at
he losing re r cattle
ted o goat
a very ding to
goat nd of the use of
e of tage of reases ges B.
emain ge to the
of the again
18 | P a g e Besides sheep and cattle there are small percentages for pigs, dogs and horses. Pigs are present in every period except the Middle Iron Age which has yielded very little remains and in the Late Iron Age. It seems that pigs were introduced at Achlum during the Roman period. Horse remains are present in most periods in fairly small percentages. Slightly higher percentages are present in the Ottonian period and the Late Middle Ages A. Dog remains were found in the Late Iron Age, Roman period and the Late Middle Ages. Whether horses and dogs were present throughout the period that Achlum was occupied is not clear. Considering the lack of remains from both species it is possible that remains from other periods were not preserved or were simply not deposited in this part of the terp.
3.2 Livestock consumption
As mentioned before the bone weight method was used to analyse the importance of life stock species to the consumption of meat of the inhabitants of Achlum. The results of this method are presented in Appendix C; a visual representation of this data is given in figure 4. The graph shows the same increase of importance of sheep from the Late Iron Age to the Carolingian period. Because this is a representation of the importance of meat produced by domesticated species, the percentage of sheep does not exceed the percentage of cattle during the Merovingian and Carolingian periods.
Apparently sheep were kept not just for their meat but also for their wool or even their milk. Cattle were clearly the most important contributor to the meat consumption of the inhabitants of Achlum.
Of course it is likely that cattle were also kept for their milk and used for traction. Evidence for the slaughter of cattle and sheep was found in all periods. Cut marks on sheep and cattle bones are present in every period. Two polled cattle skulls were smashed probably as a way to stun the animals before slaughtering them (Appendix D).
The meat weight pattern of the horse shows an increase during the Roman period. This is based on just three remains so an interpretation of this development is difficult. A second larger peak is visible in the Ottonian period which is based on just two remains. Horse is also present in the Late Iron Age, The Carolingian period, the Carolingian/Ottonian period, the Ottonian period and Late Middle Ages. It is unlikely that horses were kept solely for consumption, but cut marks were found on bones attributed to the Carolingian period and The Late Middle Ages A.
The percentage of pig is very low in the periods were pig remains were found. Only the Carolingian/Ottonian period and the Ottonian period show a slightly greater percentage of pig.
Besides meat pigs produce skins, hairs that can be used in brushes and bones that can be worked.
They also can be used to clear away food refuse. Keeping pigs can be very attractive considering the fact that they produce a large number of offspring and can be slaughtered much earlier than sheep and cattle. It is however difficult to keep pigs in a terp environment. It is much easier to herd pigs in areas with forests where pigs can forage.
Pigs are often considered as luxury animals because they do not produce milk or wool and cannot be used for traction. Luxury consumption is defined by Ervynck et al. (Ervynck et al., 2003) as:
“consumption, beyond the level of affluence, of goods that are special, limited in supply, difficult to procure or very expensive for other reasons.” A diet with much variety can be called luxurious because it will include items that are not strictly optimal in terms of the ratio of costs versus
nutritional value. Another possible characteristic of a luxury diet is the selection of the prime quality parts of an animal. According to Ervynck poor rural household will consume the whole of the animal
including a site, lu distribut E).
A the ArcG
2007), wh develop some of group. T are not e pork was the diffic slaughte D There w One dog parts of not inclu to the se specific p each of t
g prime part xury consum tion of skelet
A visual repr GIS method d
hich are base a method to the differen This means th
enough rema s considered culty of keep ered during s Dogs were m
ere traces of g bone attrib
the skeleton uded. There ettlement. Th
parts of the the periods w
Cattle Horse Sheep/G Sheep Pig Dog
12 34 56 78 109
Mea t We ig h t (%)
Ac
s like the sku mption is a po tal elements
Figure 4
resentation o developed by ed on diagra o easily creat nt skeleton e
hat if one tho ains of pig pr d luxury mea ping pigs in a special occas most likely no
f tooth mark uted to phas n where foun is no evidenc he distributio different live which indica
1a 100
Goat 100 2030 4050 6070 8090 00
chlum 20
ull. When sel ossibility (Erv
of pigs to co
4: Achlum 2009
of the distrib y Orton (Orto
ms by Pales te skeletal di lements, for oracic verteb resent to pro t at Achlum terp environ ions.
ot eaten at A ks of dogs on se 5b was bu nd during exc ce of prefere on of skeleta estock specie te that the a
1b 2
90,6 74,9 0,7 14,7 6,0 4,9 0,6 1,9 0,0 2,6 2,1 1,1
009 ‐ Me time (h
lection of sp
vynck et al., 200
onfirm a pref
9 ‐ Meat weight
bution of ske
on 2010). Orto
& Garcia (Pa
istribution d r example the bra is presen ove or dispro
is not clear.
nment it is li
Achlum, at le n two dog bo urned. As me
cavation. Par ences for spe al elements (
es. Remains animals were
4 9 66,8
7 0,0
9 26,0
9 7,0
6 0,2
1 0,0
at weigh hand‐col
ecific parts i
03). No evide ference for s
t distribution o
letal elemen on used temp
les & Garcia 19
iagrams usin e thoracic ve nt all thoracic ove the prese
Considering kely that jus
ast no cut m nes from Th entioned bef rtial skeleton ecific types o (Appendix E) of most skel e killed at the
5a 5b
62,3 77,9 4,0 6,1 23,0 8,6 9,4 2,6 1,2 4,7 0,0 0,0
ht distrib lected)
s visible in th ence was fou
specific cuts
over time
ts is given in plates create
981) and Baro ng ArcGIS. Th ertebrae are c vertebrae a ence of prim the low amo t a few pigs w
marks were fo e Late Middl fore Append ns and sieve of meat or fo showed no etal element e settlement
5c 49,1 73 34,9 11
6,5 8
4,1 2
5,3 2
0,0 1
bution ov
19 he animal re und in the
of meat (Ap
n figures 5‐7 ed by Yvinec one (Barone 19
he method co combined in are coloured me cuts. Whe ount of rema
were kept a
ound on thei le Ages (App dix E shows w
fraction rem or the import sign of selec ts are presen t.
6 7
3,3 88,4 1,7 7,4 8,3 2,1 2,7 0,7 2,5 1,0 1,4 0,5
ver
| P a g e mains of pendix
using (Yivinec 976), to ombines nto one
. There ther ains and
nd
r bones.
pendix D).
which mains are
t of meat ction of
nt in
F
Figure 5:
Figure 6: Achlum
Figure 7: A
: Achlum 2009 –
m 2009 ‐ Numb
chlum 2009 ‐ N
– Number of di
ber of different
Number of diffe
ifferent skeleta
t skeletal eleme
erent skeletal e
al elements of
ents of Pig in th
elements of Pig
Pig in the Roma
he Carolingian/
g in the Late Mi
20
an period
/Ottonian perio
iddle Ages B
| P a g e
od
21 | P a g e 3.3 Fish
Fish remains were absent in the hand‐collected material but the sieve fractions yielded a fair number of remains. A total of 120 fish remains were excavated of which 71 remains could not be identified to species. The remains were analysed using the comparative collection of the Groningen Institute of Archaeology. The results of the analysis of the fish remains from the 2 mm and 5 mm sieve fractions are shown in tables 3 and 4. The remains from both fractions were combined into one table (table 8) to analyse possible changes in fish consumption over time. Not all periods yielded fish remains; other periods like the Late Iron Age, Roman period, Merovingian period and
Carolingian/Ottonian period have yielded very few remains. The lack of fish remains in phase corresponds with the very low amount of hand‐collected remains from this period. The lack of fish remains from the Ottonian period is quite strange even considering the low amount of fish remains found at Achlum. It is however not safe to assume that fish were not eaten in these periods.
Table 8: Achlum 2009 ‐ Number of remains of fish (5 mm/ 2mm sieve fraction remains)
1b 2 4 5a 5b 6 7
gadids 2
flatfishes 4 2 17 1 44 8
cyprinids 1
herring 12 1 2
european eel 2 1 4 1 8 3
salmon/trout 1
garfish 5
three‐spined stickleback 1
Total: 2 4 3 36 2 60 13
%
1b 2 4 5a 5b 6 7
gadids 0 0 0 6 0 0 0
flatfishes 0 100 67 47 50 73 62
cyprinids 0 0 0 0 0 2 0
herring 0 0 0 33 0 2 15
european eel 100 0 33 11 50 13 23
salmon/trout 0 0 0 3 0 0 0
garfish 0 0 0 0 0 8 0
three‐spined stickleback 0 0 0 0 0 2 0
Total: 100 100 100 100 100 100 100
Table 8 shows a preference for flatfishes in all periods with the exception of the Late Iron Age. The percentages of European eel are lower but show that this species was also regularly eaten.
Herring is present in the Late Middle Ages and makes up a large part of the fish remains in the Carolingian period. The low numbers of herring in several of the periods makes it difficult to analyse if species were or were not eaten. When sufficient fish remains are present, as is the case in the Carolingian and Late Middle Ages A, the variety of species increases. It does seem that fish
consumption at Achlum mostly consisted of flatfishes and European eel. Whether herring was also a preferred fish species is unclear because only the Carolingian period yielded a large number of
22 | P a g e remains. The other species were probably not part of the main diet and it is also possible that they were not eaten at all. The three‐spined stickleback is an indication for this last interpretation because it is considered inedible.
3.4 Poultry
Only a few bird remains were identified during the analysis of the animal remains of Achlum (tables 2‐4). The Middle and Late Iron Age as well as the Ottonian period are not represented in the hand‐collected and sieve fraction remains. The bird remains from other periods are of mallard, goose and chicken. Several remains of the same bird species were found in features that could not be dated as well as a single bone of a brant goose (table 2). In the layers attributed to the Roman period a single mallard bone was found (table 3) as well as an unidentifiable bird species that could not be identified (table 4). The only other species found in the Roman period layers of Achlum is the chicken (table 3).
The Merovingian period yielded two mallard remains and one bone that could not be identified to species. A total of ten chicken remains and a single unidentifiable remain were found in Carolingian period features (table 2 and 3). Another unidentified bird bone was found in a
Carolingian/Ottonian feature (table 3). Two mallard fragments, a fragment of an unidentified goose and two unidentifiable bird bones were found in Late Middle Ages A features as well as two chicken fragments (table 2 and 3). The Late Middle Ages B layers yielded two chicken remains as well as four unidentifiable remains (table 3 and 4).
The amount of bird remains and the limited number of bird species indicate that birds were not a major part of the diet of the inhabitants of Achlum. Wild geese and mallards were hunted in some of the periods, but not intensively. The chicken was present at the settlement since the Roman period. Only the Carolingian period yielded a fair quantity of chicken remains indicating a possible increase in consumption in this period. The lack of bird remains in some of the periods cannot be explained by fragmentation and conservation problems. There were not many fragments of bird bones in the sieve fractions and the conservation on site was very good.
3.5 Molluscs
Molluscs were found in hand‐collected material and in both of the sieve fractions (table 2‐4).
Most of the remains were found in ditches and pools, only a few remains were found in pits. A number of molluscs were found in natural silt layer of which a sample has been taken. The number of remains is not very high, so molluscs were probably not a major contributor to the diet of the
inhabitants of the terp. It is possible that molluscs were not eaten at all. The problem with analysing the consumption of molluscs in a terp settlement is that the shells can be deposited in the area during floods. Another possibility is that the shells were brought to the site by the inhabitants to increase the height of the terp or to construct shell paths. It is therefore not easy to establish which molluscs were eaten at Achlum.
Some of the mollusc species were not fit for consumption like the very small widespread column. The widespread column was possibly brought to the site during the heightening of the terp with grass sods. The arctic barrel‐bubble, European stream valvata, slender amber snail and layer spire shell are also very small and not likely consumed at Achlum. The common cockle, common
23 | P a g e periwinkle, baltic macoma, peppery furrow shell and mussel are fit for consumption. The common cockle specimens are quite small so it is unlikely that they were eaten. Most of the remains of this species were found in natural deposits. Whether the other mollusc species were eaten is not clear although it is likely that the mussel was eaten.