They did not Live by Grass Alone: the Politics and Palaeoecology of Animal Fodder in the North Atlantic Region

Full text

Environmental Archaeology
1,
1998;
pp. 41-54
They did not Live by Grass Alone: the Politics and
Palaeoecology of Animal Fodder in the North Atlantic
Region
Tom Amorosi, Paul C. Buckland, Kevin
J.
Edwards, Ingrid Mainland, Tom H.
McGovern,
J
on ~ Sadler and Peter Skidmore
Introduction
It is perhaps a truism to state that the further north
the community, the more the diet is likely to be
dominated by animal products. Whilst barley
(bygg)
was grown in the Northern Isles, the Faroes and
Iceland (Fenton 1978;Fridriksson 1960),cultivation
declined in many parts of the region through the
late medieval period. Wild lyme grass
(Leymus
arenarius)
was harvested and eaten as a cereal in
parts of Iceland (Guomundsson 1996).Most Norse
Greenlanders, however, knew no bread, as a
fourteenth century source, the King's Mirror (KM
17, Larson 1917) notes, and as Erik the Red's Saga
implies (Magnusson and Palsson 1965, 92), beer
was also unknown, a factor which may have been
more important in a chieftains' society. Con-
temporary documentary sources in Iceland stress
the importance of the secondary products of milk,
cheese and whey, prepared from sheep as well as
cow's milk, in basic subsistence (Aoalsteinsson
1991),supplemented by some domestic animal meat
and fish,with seal in coastal regions (Amorosi 1996).
The Norse Greenlanders appear to have made the
most extensive use of wild resources, exploiting a
combination of seal and caribou rather than fish,
yet their settlement pattern strongly indicates the
determining role of the scattered pasture plant
communities that provided fodder for their small
herds and flocks (McGovern 1992). Faroese, Shet-
landers and Orcadians exploited varying com-
binations of domestic, marine and some terrestrial
wild resources, but secondary products from both
cows and sheep remained a key element in basic
subsistence (cf. Bigelow 1985; Aoalsteinsson 1991;
McGovern 1992).
This paper examines the sources and politics of
fodder on the North Atlantic islands, and con-
siders some of the palaeoecological signatures of
the material. Cereal growing has never been a vital
component of fodder production over much of
these northern areas and it is little considered here.
The exploitation of heathlands, doubtless of great
importance to Norse settlers (cf. Kaland 1986),
would merit a paper in its own right and is a topic
which must be neglected likewise.
Fodder and Farming Strategy in the North
Atlantic
As Iceland's first Norse settler, Hrafna-Floki, found
to his cost (If 1936, 36), the key element in any
North Atlantic farmer's survival strategy was to
obtain sufficient fodder to overwinter core stock.
Although in the past both cattle and sheep were
left outdoors throughout the year (e.g. Lindroth
1937,45), particularly in areas of heavy snowfall,
permanent byres and associated barns were, and
are, a necessity wherever the ground was covered
for long periods (Figs. 1 and 2). As the nineteenth
century life of Jon Jonsson who lived at Vogar by
Myvatn in northern Iceland (Fig. 3) clearly shows
(Jonsson 1877), the common neurosis of any Ice-
Authors' Addresses: Tom Amorosi and Tom
H.
McGovern, Department of Anthropology, Hunter College, City University of
New York,
695
Park Avenue, New York,
NY
10092, USA; Paul
C.
Buckland, Kevin J. Edwards, Ingrid Mainland and Peter
Skidmore, Department of Archaeology and Prehistory, University of Sheffield, Northgate House, West Street, Sheffield,
51
4ET, UK; Jon
P.
Sadler, School of Geography, University of Birmingham, Edgbaston, Birmingham B15 2TT, UK.

42
T. Amorosi et al.
landic farmer was, as now, the acquisition of
sufficient quantities of fodder to keep animals alive
through the winter. At the start of each winter
decisions had to be made as to how many animals
could be kept on the volume of fodder collected.
Farmers had to balance the winter fodder needs of
currently mature stock, immature animals needed
for replacement and herd expansion, human meat
requirements and dairy provisioning needs, pasture
productivity of the previous summer growing
season, and the estiInated, but still unknown
duration of the winter feeding season. This risk
assessment thus depended upon a complex of
interacting variables, a.nd not all farmers were
equally capable and lucky in their closely calculated
autumn choices (see, for example, a Faroese folktale
provided by Wylie
198?').
Any culling strategy had
to allow for the numbers essential not only to
maintain stocking levels but also to obtain sufficient
milk to provide human food. In bad years culling
might also be accompanied by the dismissal of any
hired hand who could not be fed (Jonsson
1877),
and it is not surprising that human population
growth was strictly controlled in medieval Iceland
by minimum wealth requirements for marriage and
householding (Vasey
1996;
Ingimundarson
1995).
The Politics of Fodder
Fodder was a political, as well as an economic and
environmental variable in Norse and later society
in the North Atlantic. In
Jaroab6k,
an Icelandic Land
Register compiled by .Arni Magnusson and Pall
VidaHnbetween
1702-1?12
(Magnusson and Vidalin
1943),
farmsteads were clearly rated by their 'fodder
reserve value'. Although the precise meaning of
this term is debatable, these fodder values had a
number of implications, such as the kind and quality
of fodder that a farmstead could produce in any
given year, as well as the requisite fodder to bring
a farm's stock through a winter and maintain a
household's dairy requirements.
Whilst a farmer had to guard against a late arrival
of the spring, in good years he might also be able
to trade surplus fodder with less fortunate neigh-
bours. Tenants were regularly required to provide
fodder for landlords' stock, and complex legal
strictures governed fodder transactions
(Gragas
in
Hastrup
1985).
Apparently independent free-
holding northern farmers were closely linked into
a system of co-operative interdependence, in which
the minimal unit of long term survival was not the
isolated farm, but the local community. In Iceland,
this
hreppur
consisted of
15-30
neighbouring hold-
ings, and a similar situation probably pertained in
Greenland. This comInunal association provided
Figure
1.
Exploded diagram of sheephouse of dry-stone,
turf and driftwood, Eyjafjallasveit,
S
Iceland
(1981).
Drawing: T. Addyman.
Figure
2.
Interior of dry-stone and turf-built
sheephouse, Eyjafjallasveit,
S
Iceland
(1979).
Note the central feeding trough, the depth of dung,
litter and hay in the standing areafor the animals and
the use of the roof space to store additional grass fodder.
for limited management of joint resources, partic-
ularly upland common grazing, for poor relief, and
for some sharing of stock and fodder resources.
While this local level co-operation was certainly
the source of the adaptive resilience shown by many
rural communities in the medieval North Atlantic,
these associations did not always spread risk and
benefit evenly, and tended to be dominated by
their wealthiest members. Chieftains might take
rents in fodder, but also might provide fodder relief
to retainers during periods of shortfall. The action
of the Icelandic
Hen-Thorir's Saga
turns on the
competitive efforts of chieftains to secure fodder
for retainers during a period of local shortage in
order to maintain and extend their influence and
prestige (Miller
1990, 94-100).
At the other end of
the social scale, poor smallholders had constantly

The Politics and Palaeoecology of Animal Fodder in the North Atlantic Region
GREENLAND
Figure
3.
Location map of the North Atlantic region.
43
to balance their need for community assistance with
the potential descent from freeholding status into
tenantry, a status change with profound economic
and environmental consequences. While free-
holding
b6ndar
(middle-ranking free farmers) in
medieval Iceland controlled their own household
labour and made their own decisions about stocking
levels and the mix of domesticates, tenants were
liable for labour service (especially during the short
haymaking season) and often were compelled to
board the land owners' stock. In medieval Iceland,
tenant leases were annual, and provided little long-
term security for the tenant or incentive for labour
intensive improvements such as drainage, home
field enclosure, or manuring (Hastrup 1985).
Widespread tenantry in the early modern period -
in 1695 only 47 out of a total of 4,058 farms were
owned by Icelanders (Anon 1942)- was regularly
cited as a cause for pasture degradation and low
agricultural productivity. In addition, as Halldor
Laxness' (1946) famous story of Bjartur shows, the
imposition of an extra cow upon a tenant by a
landlord may have been the breaking point for a
farmer.
Variation in elevation, slope, aspect, drainage,
erosion and soil type within a district tended to
produce recurrent patches of fodder deficit and
surplus. Variability in accumulated summer tem-
perature and in the timing and amount of annual
precipitation affecting an entire region simul-
taneously, would often produce very different
impacts on different holdings in the same district,
and this environmental patchiness had profound
long-term political consequences. The history of
Landnam and settlement in the Atlantic islands
tended to place the pockets of the richest and least
vulnerable land squarely within the homesteads of
the colonising chiefly aristocracy, providing the
early winners in the environmental/economic/
political dialectical game
(sensu
Crumley 1994)with
a substantial initial advantage (McGovern
et al,
1988).Such uneven impacts might lead to abandon-
ment of the more marginal holdings in a run of bad
years, with the dominant farmer taking over the

44 T. Amorosi et al.
hayfields of the less fortunate, and re-occupation, or
conversion of shielings
(sel, scetar,
summer farms)
into permanent sites
in
warmer periods (Svein-
bjamard6ttir 1992). In both Greenland and Iceland,
it would appear that the proportion of tenants to
free farmers increased dramatically in the later
Middle Ages, and it has been argued that cycles of
settlement expansion and contraction (as well as
increased feuding and intra-communal violence)
played a significant role in the reduction in the
number and power of the freeholding
b6ndar
from
the 13
th
century onwards (Hastrup 1985;Miller 1990;
McGovern 1992; 1994). Palaeoecological evidence
(Dugmore and Buckland 1991) indicates that these
social changes were contemporary with widespread
soil erosion and accelerated landscape change in
many areas. Evidence from Greenland indicates
similar problems with soil erosion (Fredskild 1992).
The new high resolution proxy climate data pro-
duced by the GISP2 and GRIP ice cores suggest that
increased climatic variability in the 14
th
and 15
th
centuries may also be a significant part of this
complex story (Barlow 1994; Buckland
et al. 1996;
O'Brien
et al. 1995).
The FARMP ACT IVlodel
These multi-stranded interactions of human farm-
ing strategy, political competition, landscape
evolution, stocking level, pasture productivity, and
short-term climate fluctuation produce causative
webs difficult to navigate without recourse to one
or another form of explanatory determinism. Like
many others faced with similar complexity, a
research group within NABO, the North Atlantic
Biocultural Organisation, has developed a simple
simulation model series called FARMPACT 5.0 (see
McGovern 1995for fuller description of underlying
data and assumptions, for applications see Barlow
et al.
1997). This is based upon the best studied
portion of the Norse Western Settlement in Green-
land, and allows for better investigation of the
interaction of pasture productivity, stocking mix,
household consumption, and wild products needed
for dietary supplement. Pasture productivity is a
key component of the model. All archaeological
surveys ofthe Norse settlements of Greenland have
noted the close correlation of ruins with the patches
of grass/sedge communities suitable for animal
pasture (McGovern and Jordan 1982; Christensen
1989).Basiccalculation of modelled fodder produc-
tion and stock consumption may serve to underline
the costs of a long .winter for a middle ranking
modelled farm (Fig. 4). New versions of FARM-
PACT are currently under development by the
NABO models group at the Geography Department
of the University of Edinburgh, and these seek to
integrate the MLURI land utilisation model devel-
oped by the Macaulay Institute in Aberdeen for
use in Scotland with other Atlantic island data (cf.
Simpson
et al.
in press). The present version
Pasture Consumption
by Farm
Rank
Farmpact 5.0 "Sandneshreppur"
Millions
1.6
1.4
1.2
13
"'"
'5
1.0
a-
u
~
"'"
u
0.8
"0
"'0
0
~
~
0.6
0
bb
~
0.4
0.2
0.0 "1000" "1300" "1500"
I':TH::I
Tenants
IWliWi!i1'liil
Bondar
o
Chieftain
Figure
4.
Bar chart sho'wing potential fodder productivity in the Western Settlement of Greenland.

The Politics and Palaeoecology of Animal Fodder in the North Atlantic Region
45
provides some interesting output that may serve to
stimulate discussion on the dynamic interactions
possible between fodder, pasture, social ranking
and political power.
If the legal proscriptions of the medieval Ice-
landic law codes are followed and rents and tithe
payments extracted from the modelled
hreppur
in
the proportion mandated, and if the specified
fraction is retained for poor relief, the total available
surplus in terms ofeither fodder or domestic animal
products is very modest. For a 25 farm
hreppur
(chieftain, 9
bondar,
15 tenants based on the best
known settlement unit in Greenland, centred on
the church farm at Sandnes, V51 (McGovern
et ale
1996) the maximum probable animal products
surplus (figured in sheep units, cf. Fridriksson 1972)
is only 15 human annual rations. This might be
enough to carry one or two households through a
local disaster, but would not be enough to insure
the whole community.
The vulnerable households are middle ranking,
with relatively large modelled households (based
on floor areas of excavated examples) and relatively
modest herds and flocks (based on byre floor areas).
Thesemiddle ranking
'bondar'
household economies
experience more serious shortfalls in fodder produc-
tion and human provisioning than do either the
smaller 'tenant' households or the large, but well
provided modelled' chieftain' household, when the
whole system is subjected to simulated pasture
productivity reduction. Despite these vulner-
abilities, the
modelledhreppur
shows great resilience
in the face of even severe simulated reductions in
pasture productivity. Stock reductions combined
with realistically feasible intensification of wild
resource exploitation (here caribou and seals) would
allow farms to survive, though the middle ranking
households are more severely stressed than either
the low or high ranking ones.
The model, however, also highlights problems
that farmers of all ranks would face in attempting
to respond to improved pasture productivity
following a severe season requiring significant stock
reductions. Given the time required for new born
kids, lambs, and calves to reach sexual maturity
and dairy production, farmers enjoying a series of
fine seasons after a major loss of adult stock might
experience a temporary surplus of fodder produc-
tion in a small, bounded community unable to
import stock from other regions. Jonsson's memoirs
(1877)from Iceland describe just such a situation in
the nineteenth century and document the frustration
at his inability to rapidly rebuild stock to take
advantage effectively of a favourable fodder
balance.
FARMPACT also highlights the important role
of wild species, the 'Gifts of God' in Icelandic terms,
both directly in supporting human provisioning and
in providing indirect support for the domestic
mammal component of the household economy. In
the simulation, caribou and seals, especially, act to
provide substitute meat and fat that take pressure
off a herding economy recovering from a set-back,
reducing the need to cull growing domesticates for
meat and allowing fuller diversion of milk to the
young animals in order to promote rapid growth
and early maturation. The model would thus
predict that repeated crises in the domestic stock
raising economy would promote increased use of
seals and caribou. Interestingly, available zoo-
archaeological evidence (McGovern 1992)indicates
that seal consumption increased on low and middle
ranking farms, and that caribou consumption may
have increased on high ranking farms. Recent
isotopic work on the dated human remains suggests
a general trend towards a more marine dietary
intake in later periods of the settlement (Lynnerup
1995; 1996). Additional isotopic research on both
human and animal remains from Greenland now
underway (Arneborg
et al.
in prep.) may clarify
many aspects of the inter-locking problem of human
and domestic animal diet in Greenland. A similar
project has recently been initiated in Iceland (Fricke
et ale
in press), which will also look at animal bone
and tooth wear patterns.
While the FARMPACT simulation suggests that
this model northern herding economy could have
survived even catastrophic short term reductions
in pasture productivity and stock loss, it would be
most vulnerable to closely spaced series of poor
growing seasons with long winters. Even in its
present crude state, the model indicates that the
greatest threat to the viability of the system as a
whole would not be a few exceptionally hard
seasons, but rather a concentration of poor to
moderate ones closely spaced together. The poten-
tial for the new high-resolution ice core proxy
climate record for the further investigation of
fodder production and sequences of weather on
the annual scale is only beginning to be tapped
(Buckland
et ale 1996).
Both the historical record and the results of the
FARMPACT simulation point to the same general
conclusion, that fodder production was the key
element in Norse settlement and subsistence in the
western North Atlantic.
Fodder
The term hay has been deliberately avoided in the
introduction. In most English-speaking countries,
it refers to grass-dominated swards harvested just
before seed production. The material is turned and

46
T.
Amorosi et ale
Figure
5.
Tjarnir, Eyjafjallasveit,
SW
Iceland looking
N (1979).
The view encapsulates the economy and problems
offarming in Iceland. At least three phases of occupation and abandonment, the last in
1948,
are apparent. The
farm is a double holding, indicated by the two mounds at the right centre of the photograph. The remains of various
turf hay barns and byres are scattered around the enclosure banks protecting the hayfield, which has been partly
improved by paring off frost mounds and draining. A narrow leat entering through the northern turf boundary
allowed flooding of the fields in spring to encourage fresh growth.
dried in the field before storage or, increasingly, is
stored wet as silage. Before the widespread adop-
tion of clover-grass leys, often dominated by
imported grasses, hay was a much more varied
crop, and farmers would differentiate between the
hay from the manured infield
(tun),
kept for prime
animals, principally the cows (Aoalsteinsson 1991),
and that collected from elsewhere. Hay could be
collected anywhere subject to legal limitations on
access (cf.
Grtigas
Ib, 113;Iib,478, Sveinbjamardottir
1991),and some areas were harvested on a two to
three year cycle (Aoalsteinsson 1991). The use of
shielings
(sel,seeter)
reduced the pressure on hayfield
and grazing land in some regions (cf. Albrethsen
and Keller 1986;Sveinbjamardottir 1991)and these
might also provide additional sources of fodder.
The need to recover every last blade of hay is
illustrated by Lindroth (1937, 49), who records the
dispersal of mown and raked hay on an offshore
island in eastern Iceland by a storm and the efforts
made to recover any wisp of material left in crevices.
In northern Iceland, in the steep valleys of inland
Skagafjorour, strings of ponies would be taken up
onto the high plateaux to bring back hay harvested
from any available wild stands. The species compo-
sition of such materials would vary significantly
from the hay from the infield. In the south ofIceland,
in Orrefi, lush growths of
Carex lyngbyei
were
harvested by men wading waist-deep in water,
whilst the use of leats to lead water onto fields in
the spring to generate a flush of new growth is
recorded from both early modern Iceland (Fig. 5)
and medieval Greenland (Krogh in McGovern
1992).Where available, cereal straw and threshing
waste supplemented hay. In Faroe, where barley
was more widely grown, Williamson (1948, 209,
211)notes that not only were the chaff and tailings
boiled into a 'kind of gruel' and fed to stock, but
that the grass and green weeds were recovered
from the straw for use as green fodder; the straw
itself was only fed to stock 'in hard times'. In the
late medieval period, over much of the North, and
increasingly as more labour was invested in
commercial fishing, thus leading to reductions in
cultivation, imported Scottish barley replaced local
prod uction. In the north of Iceland and in Green-
land, although cereals might be grown as green
fodder, their cultivation as a grain crop is not an
option and other sources had to be exploited. The
ethnographic record shows the widespread use of
seaweed (Fenton 1978;Krisljansson 1978).Hallson
(1964)records the use of dried
Rhodymenia palmata
(Icel.
sol)
and cooked and dried
Alaria
to feed cattle,
sheep and horses, and that it was regarded as good
milk producing and fattening fodder. Jon Jonsson
(1877), in extremis,
was forced to do what from the
palaeoecological record the Greenlanders appear
to have done routinely - as soon as new growth of
birch and willow appeared above the snow, he
collected every available piece in an abortive

The Politics and Palaeoecology of Animal Fodder in the North Atlantic Region
47
attempt to keep his stock alive until the first growth
of grass. Differences in perceived values of wood-
land are seen in northern Iceland, where in the
eighteenth century farmers in Fnj6skadalur com-
plained that woodland interfered with their hay-
making, yet at SncebjarnarstaDir in the same valley
the farmer was forced to destroy his woodland in
order to feed his sheep when the grass hay ran out
(Bjarnason 1980, 35).
Vegetable fodder was also supplemented with
animal proteins and fats. Routinely, fish and fish
offal were fed to cattle in northern Iceland. Herring
was carried inland and mixed with hay without
pre-treatment, whilst cod waste needed to be
boiled before feeding to stock. Lindroth (1937,46)
records the use of 'heads of cod and sea cat
(sic),
crushed fish bone, fish entrails, hard roe, and the
like'. Seal fat was also liquefied and poured over
hay and it is also likely that any spoiled human
food and its residues were recycled through the
domestic animals.
The Palaeoecological Evidence
Whilst both climatic parameters and the ethno-
graphic record constrain any interpretation of
medieval and later farming patterns on the North
Atlantic islands, individual sites may have differed
substantially in their subsistence strategies and it is
necessary to employ as many lines of evidence as
possible in attempting to model the past. In the
machair soils of the Northern Isles, bone preser-
vation is excellent, and a number of relevant sites
has been studied extensively (cf. Bigelow 1985).
Little has been published from Faroe, where the
soils are often less conducive to preservation,
although material from Toftanes on Streymoy is
currently under investigation (Hansen 1991;Larsen
1991).Amorosi (1996)has recently reviewed all the
animal bone material available from Iceland, and
Greenland has been similarly treated by McGovern
(1985).The general pattern of animal kill supports
a model of secondary product dependency and,
although McCormick (1992)has argued otherwise
from the Irish evidence, the North Atlantic island
bone data, taken with the overwhelming ethno-
graphic picture, are conclusive.
The animal bones
The bones and teeth of the livestock reared by the
farmers of the North Atlantic islands serve as a
source of evidence for the use of fodder. Variability
in the species composition and mortality profiles
of the archaeozoological assemblages recovered
from the Norse settlements can give a crude
indication ofthe availability and necessity offodder
and in some cases may even reflect the type of
fodder used. In Norse Greenland, the relative
importance of goats in the economy of the Western
Settlement, particularly on the smaller and pre-
sumably lower status farms, contrasts strongly with
that evident in the Eastern Settlement, where sheep
and cattle are the dominant domesticates, and this
may reflect both the presence of greater quantities
of scrub and bush vegetation in the more marginal
Western Settlement and a lack of good pasture
land for grazing/fodder production on the smaller
farms, browse being more effectively metabolised
by goats than sheep or cattle (McGovern 1989).
Similarly, at SvalbarD in northern Iceland, an
increasing emphasis on caprines during the later
Middle Ages has been interpreted as evidence for
a decline in pasture quality (Amorosi 1992). In
several North Atlantic assemblages, mortality
profiles which emphasise the slaughter of neonatal
and juvenile domesticates have been cited as
evidence for inadequate supplies of fodder for
overwintering (Rackham
et al.
1996;Noddle 1977),
although such profiles could equally reflect dairying
activities (Rowley-Conwy 1983;Halstead in press).
A clear case of inadequate diet resulting in high
infant mortality is, however, provided by the 17/
18th century levels at SvalbarD, where a high
frequency of neonatal/postnatal deaths in lambs
can be attributed to the detrimental effects of spring
and summer sea-ice on pastureland (Amorosi 1992).
Finally, where species representation and mortality
profiles indicate a change in the economics of animal
husbandry towards systems which require a high
quality diet, such as the intensification in dairy
prod uction (cattle) suggested for Late Norse
Shetland (Bigelow 1992), or the increased use of
cattle for traction evident during the Pictish-Norse
interface and Norse phases at Pool in Orkney
(Hunter 1997),itcan be argued that an improvement
in fodder production must also have occurred,
perhaps through intensification of arable pro-
duction or the adoption of hitherto under-utilised
land for fodder/grazing.
Archaeozoological evidence can also be used to
give a more detailed insight into fodder use
through approaches such as the analysis of bone
chemistry (e.g. d'Angela 1991;Ezzo 1992),and the
examination of ungulate dentition (e.g. Guthrie
1992; Dobney and Brothwell 1988; Mainland this
volume). Amber's (1994)use of carbon isotopes to
identify seaweed-eating sheep in Neolithic and
Bronze age contexts from Orkney demonstrates the
potential of this technique for exploring the use of
seaweed and other marine resources as fodder in
the North Atlantic islands and research in progress
is applying isotopic analyses to the question of

48
T. Amorosi et ale
livestock diet in Norse contexts from both Iceland
(Fricke
et
ale
in press) and Greenland (Arneborg
et
ale
in prep.). This research, together with work cur-
rently underway on the identification of grazing,
browsing and hay-fed (grass-hay and leafy-hay)
animals in Norse Greenland, Iceland and Orkney
using tooth wear analysis (Mainland in prep.; this
volume) will help identify the use of different types
of fodder in the North Atlantic islands, and more
importantly, will allow some insight into how the
available resources were used for the different
species and categories of animal (Mainland this
volume).
Plant macrofossils
Direct evidence for the nature of the hay fodder is
currently limited. In the eastern Atlantic this is
largely preservational in that few sites have organic
preservation and what there is can rarely be directly
related to what was being fed to animals. Farm
mounds in Iceland, particularly in the south (cf.
Bucklandet
ale
1994),
are often sufficiently anaerobic
to preserve both building turf and organic sedi-
ments in their middens, but it is usually difficult if
not impossible to sort out turf from discarded hay
(Fig. 2) or regrowth on abandoned parts of the
midden. In addition, as a result of the timing of
harvests, hayfields are lnore likely to be recognised
by their associated weed flora than by the seeds of
the grasses and sedges themselves.
A cursory examination of plant macrofossils
during the process of recovering insect remains
from numerous samples from St6raborg in southern
Iceland suggests a significant preservational bias
towards the seeds of sedges (Cyperaceae) as
opposed to grasses (Poaceae). In a similar study of
material extracted directly from a feeding trough
at the newly excavated farm of GUS in the Norse
Western Settlement in Greenland (Andreasen and
Arneborg
1992;
Arneborg and Berglund
1993),
where the permafrost has ensured almost total
preservation, the assemblage is dominated by
Caryophyllaceae (P.I.Buckland,
J.
Ross,
in progress).
The taphonomy is difficult - does this reflect the
feeding of chickweed to the animals, or is it the
residue from grass hay as the stock consumed the
feed? Fresh chickweed. can be damaging to sheep
and cattle, but it is acceptable dry fodder. Modem
mechanised harvesting in Iceland often leaves the
surface of well-manured fields literally covered in
the seeds of
Stellaria media
and
Polygonum aviculare
(agg.), taxa also present in the Western Settlement
samples (McGovern
et
ale
1983).
Collection by small
long-handled sickle, as most of the Greenlandic
and Icelandic material must have been (cf. Guo-
mundsson
1996,
Fig. 2), is less likely to have
introduced the seeds unless the plants were being
deliberately collected. High seed productivity
among the Caryophyllaceae andPolygonaceae
further complicates the issue, but it is tempting to
suggest either routine collection of additional
fodder from the nutrient-rich surface of an over-
grown midden, or declining quality of infield fodder
where weed species have colonised bare patches.
The latter is also hinted at on the adjacent site at
Nipaatsoq (V54), where an anthomyiid fly,
Delia
jabricii,
whose larvae feed on grasses of the genus
Festuca,
present in most other samples, is absent
from the final phase of the site (Skidmore
1996
and
in Buckland
et ale
1996),
perhaps indicating a
reduction in available grass fodder. Such hypo-
theses can only be substantiated by further meticu-
lous work upon samples (currently being under-
taken in Edmonton by
J.
Ross). Employing the
ethnographic record, the archaeological evidence
of structure type, and the fossil record from well-
preserved sites, it should eventually be possible to
ascertain what types of fodder were being fed to
what animals.
The high frequency of bud scales in all samples
from sites in the Western Settlement of Greenland
suggests that birch twigs were routinely fed to
stock and this is supported by a
Betula
pollen count
of
,...,98%
in a sheep dung pellet from Sandnes. This
pattern is not evident in Icelandic plant macrofossil
samples and was not an option in Faroe, domin-
ated by oceanic heath (but see Mahler
1991
for the
presence of birch), or in the deforested landscape
of the Northern and Western Isles of Scotland.
Invertebrate fauna
The quality of preservation in the Western Settle-
ment of Greenland, and the virtual absence of
disturbance since the demise of the settlements
during the fourteenth century (Buckland
et ale1996),
allows the testing of models to a degree almost
impossible elsewhere. On many sites where condi-
tions are sufficiently anaerobic for the preservation
of a significant portion of the biota, Poaceae are
poorly preserved and a different approach to the
nature of the hay is required, although this only
becomes possible with a quantified parallel study
of the modern flora and fauna. The robust chitinous
exoskeletons of insects, particularly the Coleoptera
(beetles) and many Diptera (true flies), are often
almost as abundant as seeds in archaeological
contexts and their ecology is sufficiently well known
for them to provide one of the most effective means
of reconstructing past natural and archaeological
environments (Buckland and Coope
1991).
Analysis
of over one hundred samples from the late and
post-medieval farm mound at St6rahorg in southern

The Politics and Palaeoecology of Animal Fodder in the North Atlantic Region
49
Iceland, where the farm was finally abandoned in
1834(Sncesd6ttir 1991),provided sufficient numbers
to allow a comparison with the present fauna,
extensively sampled by Dugmore (Buckland
et al.
1991;Sadler and Dugmore 1995)and it has become
apparent that the assemblages are radically differ-
ent. A catopid beetle,
Catops fuliginosus (=borealis
of
Olafsson 1991) often dominated samples in which
there was other evidence for mouldy vegetation. In
some cases, the numbers of feeders on mould fungi,
cryptophagid beetles of the genera
Cryptophagus
and
Atomaria,
the mycetophagid
Typhaea stercorea,
and the lathridiids
Lathridius minutus
(grp.) and
Corticaria elongata
with their associated predators,
are sufficient to imply the presence of hay residues,
although such interpretation would not be possible
in the absence of supporting purely archaeological
information.
Catops fuliginosus
is a rare animal in
Iceland at the present day with only a few captures
in the south of the country (Lindroth
et al. 1973;
Buckland 1988), yet numbers in the St6raborg
samples are often abundant, with one sample in
excess of two hundred individuals. Other quan-
titative differences are also apparent.
T. stercorea
is
less common in stored hay in the region at the
present day, and faunas tend to be dominated by
Cryptophagus
spp., along with
Lathridius minutus,
a
taxon ubiquitous in mouldy synanthropic situ-
ations. C.
scanicus
appears to dominate modern
haybarn faunas, and even allowing for the diffi-
culties in identifying fossils in this genus, it was
much less common in the past. Whilst one lot of
hay may look like any other superficially, something
has clearly happened to change beetle faunal
composition in the recent past. Although
Catops
fuliginosus
is probably predatory on the immature
stages of Diptera, it is also tied into the fungal food
chain and other members of the family are known
to be equipped to disperse fungal spores (Crowson
1982).The baseline would appear to be changes in
the nature of the decay process during storage. The
industrialisation of Icelandic farming has led to the
improvement of hayfields to the extent that ir-
regular, poorly drained fields, which were once a
riot of colours from their diverse flora, have been
almost wholly replaced by regular well-drained
monocultures of
Phleum pratensis
or similar fast
growing, high-yield grasses. Modem hay contains
little variety, and the large number of Cyperaceae
seeds in the fossil assemblages need not be wholly
the result of a preservational bias. Stefan Aoal-
steinsson (pers. comm.) has pointed out that the
hand-scythed hay of the recent past, cut on the
manured
tun
and supplemented by material col-
lected in locations often remote from the farm,
decays at a lower temperature than modem mater-
ial. This would be sufficient to alter the mix of the
Coleopterous fauna. There are suggestions of
similar changes elsewhere in the Atlantic. At
Toftanes in Faroe, where turf from the early Norse
infield was incorporated in structures, the fauna
includes taxa which are either very rare or are not
currently recorded in the islands (Buckland un-
publ.); the modem synanthropic fauna, however,
isundercolleeted (Bengtson 1981),and no quantified
data are currently available.
In Greenland, the biogeographical accidents of
introduction (Sadler 1991a) have led to a reduced
hay fauna, several elements of which became
extinct with their hosts. Bacher and Fredskild
(1993) have pointed out, however, that the lath-
ridiid
Corticaria linearis,
which is likely to be the
ecological equivalent of
C.. elongata
in Iceland, is
present in the second millennium
Be
palaeo-
eskimo midden at Qeqertasussak and, despite its
modern relative rarity (Bacher 1988), it must be
part of the indigenous fauna which moved into
the new farm habitats during the eleventh century.
The remainder of the hay fauna largely overlaps
with that of Iceland (cf. Buckland 1988; Sadler
1991a).
Although study of surviving buildings (Olafsson
1982; Nilsson 1943) and archaeological interpret-
ation of stall structures does provide some clues
(Bigelow 1987), in view of the ethnographic evi-
dence for a hierarchy of animals to which hay was
fed it would be useful to have independent evi-
dence for which species were in particular byres.
It cannot be assumed that the butchered remains
of animals in a structure provide any information
as to what was habitually kept there. Dung,
although preserved, is not always identifiable to
species, but it is probable that associated endo-
parasite ova will eventually allow this on sites with
good preservation. Currently, only sheep can be
easily distinguished in that their apparently heavy
ectoparasite load, of the ked
Melophagus ovinus
and
the fleece louse
Damalinia ovis,
provides a clear
indication of presence. Unfortunately, these ani-
mals reside in the fleece and large numbers are
likely to reflect wool processing rather than the
immediate presence of the host (Buckland and
Perry 1989). In Greenland, human ectoparasites,
both the flea
Pulex irritans
and the less easily shed
lice
Pediculus humanus,
are present in the byre
samples at GUS, and these may imply that the
herder lived with his charges; the weaker animals
would have needed frequent hand-feeding, partic-
ularly with the less palatable elements in the
fodder.
Seaweed rarely survives in a recognisable form
in the archaeological record, and where burnt kelp
does occur it is likely to reflect activities other than
collection for animal feed (Buckland
et
ale
in prep.;

50
T.
Amorosi et ale
Fenton 1986). In preliIninary samples from Tung-
meralik, an inland farrn in the Western Settlement
in Greenland, Skidmore (1996) noted a high fre-
quency of the puparia of the seaweed fly
Fucellia
sp.. Species ofthis genus spend their entire lifecycle
in the littoral zone and the puparia clearly indicate
the transport of significant quantities from the coast
at least 15 km distant. Its use is likely to have been
as supplementary fod.der. More frequently pre-
served is the exoskeleton of small colonial hydroids
which are epizootic on seaweed. The most com-
monly found, present on sites from Tuquoy on
Orkney (Sadler 1991b) through coastal sites in
Iceland to west Gre1enland, is the sertulariid
Dynamaena pumilo,
but a second species has been
noted amongst the abundant material from Nesstofa
in Iceland (Amorosi
et al.
1994;E. 6lafsson, pers.
comm.). Experiment has shown that the hydroid
exoskeleton passes virtually unaltered through the
gut of sheep when all Inacroscopic trace of seaweed
has been lost (Sadler unpubl.), and examples are
likely to occur wherever domestic stock strayed.
The fossil invertebrate record provides no
recognised evidence for the use of fish in fodder,
but the extensive sieving programme instigated at
Sandnes and the nearby low status farm at Niaqussat
(Sadler 1987) demonstrated the virtual absence of
marine fish from the Greenlandic Norse human
diet; only three small vertebrae were recovered.
The marine component of the diet has to have come
largely from the consu:mption of seal meat. Experi-
ments by Jones (1986) have noted the impact of
carnivore digestive systems on fish remains, but
similar studies have not been attempted with other
domestic animals. Fish bone is abundant in Norse
and later medieval middens throughout the rest of
the North Atlantic region, extending to sites well
inland in Iceland (Anlorosi 1989; 1991), and the
ethnographic record clearly indicates the feeding of
both herring and cod to cattle until recently
(Lindroth 1937),when it has largely been replaced
by imported hay in bad years. This is evidently an
area where more experimental archaeology is
required.
Pollen
To what extent can palynology aid the study of
fodder? For .Iceland, l-Iallsd6ttir (1987, 37) feared
that 'the floristics of the disturbed habitats, created
by man's activity and his livestock, are not fully
elucidated and probably never will become so, as
we can not find any precise analogies to the
vegetation types existing at the times of the
settlement'. Her view reflects a recognisable diffi-
culty, but is perhaps overly-pessimistic given a
growing appreciation of landuse-pollen relation-
ships (cf. Hicks 1988)and given more information
on plant macrofossils (cf. Fredskild 1983), fauna
(Buckland
et ale
1991) and agricultural history
(Sveinbjarnard6ttir 1992).
Most pollen-analytical investigations in the
North Atlantic area are based on the evidence from
off-site lake and peat sites. These provide excellent,
continuous information concerning broad-scale
vegetational and environmental change, but a lack
of spatial proximity to farmed and settled areas
inevitably leads to a dilution of any anthropogenic
signal contained in the pollen record. A range of
remedies is at hand though, including the need to
select suitable sites close to or within farmed areas,
to investigate more archaeological context de-
posits, and to be aware of the current taxonomic
barriers to floristic inference (cf.Edwards in press).
This is not the place to examine these important
methodological issues. It is possible, however, to
infer palynological responses to the use of fodder.
Thus, winter fodder collection of woody plants
may have contributed to the reduction or dis-
appearance of the pollen of birch
(Betula)
in Iceland
and Greenland, and juniper
(Juniperus)
in the
Faroes (cf. Fredskild 1983; 1988; J6hansen 1985;
Hallsd6ttir 1987;Larsen 1991).Birch is also present
in Faroe around the shieling site at Argisbrekka on
Esturoy until shortly after Landnam (Mahler 1991).
Did willow, perhaps along with birch, figure in
early woodland management practices - perhaps
to be compared with more recent leaf-hay col-
lection of woolly willow
(Salix lanata)
in Iceland
(Stefansson 1948; Kristinsson 1987) and birch in
Norway (Austad 1988)?In South Yorkshire, some
farmers provided willow and birch twigs as winter
feed to cattle as late as the 1940's, as grass hay was
regarded as having 'insufficient roughage'. If
woodland management extended to pollarding,
then that would also reduce flowering; shredding,
however, where adventitious side twigs are re-
moved, could lead to bursts of canopy flowering
and high pollen production (cf. Andersen 1988).
Such practices in the typical low density woodland
seen in high latitude areas, which is likely to have
been the pattern from
Landnam
times, may be
barely detectable in the pollen record unless pollen
sites are optimally located.
Changes in the herbaceous pollen component
are likely to be of particular interest. Sudden
expansions (and, conversely, falls) in grass-sedge
assemblages could denote the management of
either infield or more distant areas for fodder
production. Off-site pollen cores, though, tend to
be dominated heavily by one or both of Poaceae
and Cyperaceae pollen (e.g. Sardlup qaqa, Green-
land (Fredskild 1983);Mosfell, Iceland (Hallsd6ttir
1987); Hovi B, Faroes (J6hansen 1985); and Mur-

The Politics and Palaeoecology of Animal Fodder in the North Atlantic Region
51
Figure
6.
3m high turf-built hay barn attached to a
sheep house at Skatastaoir, Skagafjorour,
N
Iceland
(1981).
raster, Shetland (Johansen 1975)) and they can
hardly be said to provide reassuring evidence of
particular
landuse practices. The inability to split
pollen within the family groups of Poaceae and
Cyperaceae makes it difficult to infer that species
such as
Poa pratensis
and
Festuca rubra
amongst the
grasses, and sedge taxa such as
Carex lyngbyei
or
Kobresia myosuroides,
are under consideration. In
any case, appropriate improvements in pollen
taxonomy would still necessitate careful inter-
pretation - taxa such as
Molinia caerulea
and
Carex
nigra
can be common in grazed fens because sheep
do not find them palatable (Johansen 1985,66;and
cf. Fredskild 1988, 389).
Hayfield floras, especially where infested with
weeds, may be identifiable in the pollen record,
especially where identification to genus or pollen
taxonomic 'type' (Moore
et al.
1991;Bennett 1994)
'works' because various species, with similar pollen
morphologies, are absent from the flora. Thus,
Lychnis flos-cuculi, Prunella vulgaris, Rhinanthus minor,
Rumex acetosa, Ranunculus acris, Caltha palustris,
Trifolium repens, Bellis perennis
and
Filipendula
ulmaria,
could be useful indicators of pastures and
hayfields capable of supplying fodder (cf.J6hansen
1985, 22; Scott and Palmer 1987, 12). The practices
of building in turf (Fig. 6) and turf thatching of
house roofs, which are regularly cut and used as
fodder, might also be identifiable in the pollen and,
even more so, in the plant macrofossil records.
Species found on roofs include
Anthoxanthum
odoratum, Festuca rubra
and
Poa trivialis
as typical
dominants;
Galium saxatile
and
Silene dioica
are also
common (Johansen 1985).In on-site work, the large
pollen grains of
Leymus arenariusneed
to be carefully
discriminated from cereals.
If pollen studies are to make a useful contribution
to studies of fodder, then there is a need for modem
analogue studies (cf. Hicks 1988) supported by
appropriate statistical procedures (e.g. Berglund
et
al.
1986).Focused plant macrofossil studies would
provide the species indications lacking in pollen
investigations and they could lead to more secure
attributions for the pollen-based evidence. A
combined investigation of pollen, plant macrofossil
and faunal evidence from archaeological contexts,
could provide major advances in understanding
the role and composition of fodder in the farming
regimes of northern communities.
Conclusion
While fodder production, storage, exchange, and
consumption is only one pathway to be traced in
the complex interaction ofhuman intent, landscape
change, and environmental fluctuation in the
medieval and early modern North Atlantic region,
it appears to be a particularly important one. Both
current palaeoeconomic assessment and the re-
corded judgements of medieval and modern nor-
thern farmers agree that fodder is a key to the
understanding of northern agriculture. Further
research, combining more effective modelling,
documentary and ethnographic work, with field
study of soils and site location, and laboratory
analyses of recovered botanical and zooarchae-
ological samples is essential. Ideally, future re-
search should involve not only archaeologists and
palaeoecologists, but should also draw upon the
accumulated experience of northern farmers who
still grapple every season with the ancient problem
of fodder production and risk assessment.
Acknowledgements
This research summarises some of the results of
an international research project on human impact
on the North Atlantic islands funded largely
through NABO, the North Atlantic Biocultural
Organisation, by NSF, Office of Polar Programs,
USA. Additional funding was provided by the
Leverhulme Trust, UK. The authors are grateful to
Jette Arneborg, Philip Buckland, Stefan Stumman
Hansen and Julie Ross for access to unpublished
data and for the comments of Guorun Svein-
bjarnardottir and Paul Halstead.
References
Aoalsteinsson, S. 1991. Sheep in early Icelandic agriculture.
Acta
Archaeologica
61, 285-91.
Albrethsen, S. E. and Keller, K. 1986. The use of saeter in
medieval Norse farming in Greenland.
Arctic Anthropology
23, 91-107.

52
T. Amorosi et al.
Ambers, J. 1994. Stable carbon isotope values for Tofts Ness,
pp. 125-7 in Luff, R. and Rowley-Conwy, P. (eds.),
Whither
Environmental Archaeology?
Oxford: Oxbow Books.
Amorosi, T. 1989. Contributions to the zooarchaeology of
Iceland: some prelimina.ry notes, pp. 203-27 in Durren-
burger, P. and Palsson, G. (eds.),
The Anthropology of Iceland.
Iowa: University of Iowa Press.
Amorosi, T. 1991. Icelandic aJrchaeofauna: a preliminary review.
Acta Archaeologica
61, 272.-84.
Amorosi, T. 1992. Climate impact and human response in
northeast Iceland: archaeological investigations at Svalbard,
1986-1988, pp. 103-38 in Morris, C. D. and Rackham, D.
J.
(eds.),
Norse and Later Settlement and Subsistence in the North
Atlantic.
Glasgow: University of Glasgow.
Amorosi, T. 1996.
Icelandic Zooarchaeology: New_ Data applied to
Issues of Historical Ecology.
Unpublished Ph.D. thesis, City
University of New York.
Amorosi, T., Buckland, P. C., Magnusson, K., McGovern, T. H.
and Sadler,
J.
P. 1994. An archaeozoological investigation of
the midden at Nesstofa, Reykjavik, Iceland, pp. 69-80 in
Luff, R. and Rowley-Conwy, P. (eds.),
Whither Environmental
Archaeology?
Oxford: Oxbow Books.
Andersen, S. T. 1988. Changes in agricultural practices in the
Holocene indicated in a pollen diagram from a small hollow
in Denmark, pp. 289-99 in Birks, H. H., Birks, H. J. B.,Kaland,
P. E. and Moe, D. (eds.),
The Cultural Landscape-Past, Present
and Future.
Cambridge: Cambridge University Press.
Andreasen, C. and Arneborg,. J. 1992. Garden under Sandet: Nye
Nordb0ers0gelser i Vesterbygden.
Grenland Kultur og Sam-
funds Forsekning
92, 11-49.
d' Angela, D. 1991. Palaeoclimatic conditions in the Po plain
during the Neolithic, Chalco lithic and Bronze Ages: first
results.
Rivista di Studi Liguri
57, 119-26.
Anon 1942.
Iceland.
Andover: Naval Intelligence Division.
Arneborg, J. and Berglund, J. 1993. Garden under Sandet.
Forsekning i Grenland/I'usaat
4, 7-19.
Austad, 1.1988. Tree pollarding in western Norway, pp. 11-29
in Birks, H. H., Birks, H. J. B., Kaland, P. E. and Moe, D.
(eds.),
The Cultural Landscape - Past, Present and Future.
Cambridge: Cambridge University Press
Barlow, L. K. 1994.
EvaluCltion of Seasonal to Decadal Scale
Deuterium and Deuterium Excess Signals, GISP2 Ice Core,
Summit, Greenland, A.D. 1270-1985.
Unpublished Ph.D.
thesis, University of Colorado.
Barlow, L. K., Sadler, J. P., Ogilvie, A., Buckland, P. C., Amorosi,
T., Ingimundarson, J. H., Skidmore, P., Dugmore, A. J. and
McGovern, T. H. 1997. Ice core and environmental evidence
for the end of Norse Greenland.
The Holocene
7,489-99.
Bengtson, S.-A. 1981. Terrestrial invertebrates of the Faroe Isles
3. Beetle (Coleoptera) checklist, distribution and habitats.
Fauna Norvegica
B28, 52-82.
Bennett, K. D. 1994.
Annotated Catalogue of Pollen and Pteridophyte
Spore Types of the British Isles.
Cambridge: Department of
Plant Sciences, University of Cambridge.
Berglund, B. E., Emanuelsson, U., Persson, S. and Persson, T.
1986. Pollen/vegetation relationships in grazed and mowed
plant communities of South Sweden, pp. 37-51 in Behre, K.-
E. (ed.),
Anthropogenic Indicators in Pollen Diagrams.
Rotter-
dam: Balkema.
Bigelow, G. F. 1985. Sandwick, Unst and Late Norse Shetland
economy, pp. 95-127 in Smith, B. (ed.),
Shetland Archaeology.
Lerwick: Shetland Times.
Bigelow, G. F. 1987. Domestic architecture in medieval Shetland.
Review of Scottish Culture
3,23-38.
Bigelow, G. F. 1992. Issues and prospects in Shetland Norse
archaeology, pp. 9-32 in Morris, C. D. and Rackham, D.
J.
(eds.),
Norse and Later Seftlement and Subsistence in the North
Atlantic.
Glasgow: University of Glasgow.
Bjarnason, A. H. 1980. The history of woodland in Fnj6skadalur.
Acta Phytogeographica Suecica
68, 31-42.
Bacher, J. 1988. The Coleoptera of Greenland.
Meddelelser om
Grenland, Bioscience 26.
Bacher, J. and Fredskild, B. 1993. Plant and arthropod remains
from the palaeo-Eskimo site on Qeqertasussuk, West Green-
land.
Meddelelser om Grenland, Geoscience 30.
Buckland, P. C. 1988. North Atlantic faunal connections -
introduction or endemics?
Entomologica Scandinavica Supple-
ment
32, 7-29.
Buckland, P. C., Amorosi, T., Barlow, 1. K., Dugmore, A. J.,
Mayewski, P. A., McGovern, T. H., Ogilvie, A. E. J., Sadler, J.
P. and Skidmore, P. 1996. Bioarchaeological and climato-
logical evidence for the fate of Norse farmers in medieval
Greenland.
Antiquity
70,88-96.
Buckland, P. C. and Coope, G. R. 1991.
A Bibliography and
Literature Review of Quaternary Entomology.
Sheffield: J. Collis,
University of Sheffield.
Buckland, P., Dugmore, A. and Sadler, J. 1991. Faunal change or
taphonomic problem? A comparison of modern and fossil
insect faunas from south east Iceland, pp. 127-46 in Maizels,
J.
K. and Caseldine, C. (eds.),
Environmental Change in Iceland:
Past and Present.
Dordrecht: Kluwer.
Buckland, P. C., Gerrard, A. J., Sadler, J. P. and Sveinbjar-
nard6ttir, G. 1994. Farmers, farm mounds and environmental
change, in StOtter, J. and Wilhelm, F. (eds.), Environmental
Change in Iceland.
Miinchener Geographische Abhandlungen
B12,7-29.
Buckland, P. C. and Perry, D. W. 1989. Ectoparasites of sheep
from St6raborg, Iceland and their interpretation. Piss,
parasites and people, a palaeoecological perspective.
Hikuin
IS, 37-46.
Christensen, K. M. 1989. En unders0gelse af den norr0ne areal-
og ressourceeudnyttelse i Vesterbygden i Gr0nland.
Hikuin
IS, 7-26.
Crowson, R. 1982.
The Biology of the Coleoptera.
London: Aca-
demic Press.
Crumley, C. 1. 1994.
Historical Ecology: Cultural Knowledge and
Changing Landscapes.
Santa Fe: School of American Research
Press.
Dabney, K. and Brothwell, D. 1988. A scanning electron micro-
scope study of archaeological dental calculus, pp. 372-85 in
Olsen, S. (ed.),
Scanning Electron Microscopy in Archaeology
(BAR International Series 452). Oxford: British Archaeo-
logical Reports.
Dugmore, A. J. and Buckland, P. C. 1991. Tephrochronology
and late Holocene soil erosion in South Iceland, pp. 147-60
in Maizels, J. K. and Caseldine, C. (eds.),
Environmental
Change in Iceland: Past and Present.
Dordrecht: Kluwer.
Edwards, K. J. in press. Detection of human impact on the
natural environment: palynological views, in Bayley,
J.
(ed.),
Science in Archaeology: an Agenda for the Future.
London:
English Heritage.
Ezzo, J. A. 1992. A test of diet versus diagenesis at Ventana
Cave, Arizona.
Journal of Archaeological Science
19, 23-7.
Fenton, A. 1978.
The Northern Isles: Orkney and Shetland.
Edin-
burgh: John Donald.
Fenton, A. 1986. Seaweed manure, pp. 48-82 in Fenton, A.,
The
Shape of the Past
2. Edinburgh: John Donald.
Fredskild, B. 1983. The Holocene vegetational development of
the Godthabsfjord area, West Greenland.
,Meddelelser om
Grenland, Geoscience
10, 1-28. '
Fredskild, B. 1988. Agriculture in a marginal area - south
Greenland from the Norse Landnam (985 A.D.) to the
present (1985 A.D.), pp. 381-93 in Birks, H. H., Birks, H. J. B.,
Kaland, P. E. and Moe, D. (eds.),
The Cultural Landscape -
Past, Present and Future.
Cambridge: Cambridge University
Press.

The Politics and Palaeoecology of Animal Fodder in the North Atlantic Region
53
Fredskild, B. 1992. Erosion and vegetational changes in South
Greenland caused by agriculture.
Geografisk Tidsskrift
92,14-
21.
Fricke, H., O'Neil, J. R. and Lynnerup, N. in press. Oxygen
isotope composition of medieval human tooth enamel from
the North Atlantic as a record of climate.
Frioriksson, S. 1960. Jurtaleifar fni Bergp6rshvoli a soguOld.
Arb6k hins fslenzka Fornleifafelags,
64-75.
Fridriksson, S. 1972. Grass and grass utilisation in Iceland.
Ecology
53, 785-96. /
Guomundsson, G. 1996. Gathering and processing of lyme-grass
(Elymus arenarius)
in Iceland: an ethnohistorical account.
Vegetation History and Archaeobotany
5, 13-23.
Guthrie, R. D. 1992. New paleoecological and paleoethological
information on the extinct helmeted muskoxen from Alaska.
Annales Zoologica Fennici
28, 175-86.
Hallsd6ttir, M. 1987.
Pollen Analytical Studies of Human Influence
on Vegetation in Relation to the Landnam Tephra Layer in
Southwest Iceland
(LUNDQUA Thesis 18). Lund: Department
of Quaternary Ecology, Lund University.
Hallson, S. V. 1964. The uses of seaweeds in Iceland, pp. 398-
405 in
Comptes Rendues du
4
Congres International des Sigues
Marines, Biarritz
1961. New York: Pergamon.
Halstead, P. in press. Mortality models and milking: problems
of uniformitarianism, optimality and equifinality recon-
sidered.
Anthropozoologica.
Hansen, S. S. 1991. Toftanes: a Faroese Viking age farmstead
from the 9-10th centuries A.D.
Acta Archaeologica
61, 44-53.
Hastrup, K. 1985.
Culture and History in Medieval Iceland: an
Anthropological Analysis of Structure and Change.
Oxford:
Clarendon Press.
Hicks, S. 1988. The representation of different farming practices
in pollen diagrams from northern Finland, pp. 189-207 in
Birks, H. H., Birks, H. J. B., Kaland, P. E. and Moe, D. (eds.),
The Cultural Landscape - Past, Present and Future.
Cambridge:
Cambridge University Press.
Hunter, J. R. 1997. The early Norse period, pp. 241-54 in
Edwards, K. J. and Ralston, I. B. M. (eds.),
Scotland: Environ-
ment and Archaeology, 8000BC-AD1000.
Chichester: Wiley.
If 1936.
fslenzk Fornrit
I. Copenhagen.
Ingimundarson, J. H. 1995.
Of Sagas and Sheep: toward a Historical
Anthropology of Social Change and Production for Market,
Subsistence and Tribute in Early Iceland (lOth to the 13th Century).
Unpublished Ph.D. thesis, University of Arizona, Tucson.
J6hansen, J. 1975. Pollen diagrams from the Shetland and Faroe
Islands.
New Phytologist
75, 369-87.
J6hansen, J. 1985.
Studies in the Vegetational History of the Faroe
and Shetland Islands
(Annales Societatis Scientiarum Faero-
ensis Supplementum 11). T6rshavn.
Jones, A. K. G. 1986. Fish bone survival in the digestive systems
of the pig, dog and man: some experiments, pp. 53-61 in
Brinkhuizen, D. C. and Clason, A. T. (eds.),
Fish and Archae-
ology: Studies in Osteometry, Taphonomy, Seasonality and
Fishing Methods
(BAR International Series 294). Oxford:
British Archaeological Reports.
J6nsson, J. 1877. J6n J6nsson's saga: the genuine autobiography
of a modern Icelander (edited by Cole, G. R. F.).
Fraser's
Magazine
(new series) 15(85).
Kaland, P. E. 1986. The origin and management of Norwegian
coastal heaths as reflected by pollen analysis, pp. 19-33 in
Behre, K.-E. (ed.),
Anthropogenic Indicators in Pollen Diagrams.
Rotterdam: Balkema.
Kristinsson, H. 1987.
A Guide to the Flowering Plants and Ferns of
Iceland.
Reykjavik: Orn og Orlygur.
Kristjansson, H. 1978.
fslenzkir Sjavarhc£ttir.
Reykjavik: B6-
kautgafa Menningardsj6os.
Larsen, A.-C. 1991. Norsemen's use of juniper in Viking age
Faroe Islands.
Acta Archaeologica
61,54-9.
Larson, L. M. 1917.
The King's Mirror (Speculum Regale - Konungs
Skuggsja).
New York: Twayne.
Laxness, H. 1946.
Independent People. An Epic.
New York: A.
Knopf.
Lindroth, C. H., Andersson, H., B6dvarsson, H. and Richter, S.
H. 1973. Surtsey, Iceland - the development of a new fauna.
1963-1970. Terrestrial invertebrates.
Entomologica Scandi-
navica Supplement 5.
Lindroth, H. 1937.
Iceland, a Land of Contrasts.
Princeton:
Princeton University Press.
Lynnerup, N. 1995.
The Greenland Norse: a Biological Anthropologi-
cal Study.
Unpublished Ph.D. thesis, University of Copen-
hagen.
Lynnerup, N. 1996. Paleodemography of the Greenland Norse.
Arctic Anthropology
33, 122-36.
McCormick, F. 1992. Early faunal evidence for dairying.
Oxford
Journal of Archaeology
11, 201-9.
McGovern, T. H. 1985. Contributions to the paleoeconomy of
Norse Greenland.
Acta Archaeologica
54, 73-122.
McGovern, T. H. 1989. A comparison of the Greenlandic Eastern
and Western Settlement.
Hikuin
15,27-36.
McGovern, T. H. 1992. Bones, buildings and boundaries: paleo-
economic approaches to Norse Greenland, pp. 167-86 in
Morris, C. and Rackham, D. J. (eds.),
Norse and Later Settle-
ment and Subsistence in the North Atlantic.
Glasgow: University
of Glasgow.
McGovern, T. H. 1994. Management for extinction in Norse
Greenland, pp. 127-54 in Crumley, C.
1.
(ed.),
Historical
Ecology: Cultural Knowledge and Changing Landscapes.
Santa
Fe: School of American Research Press.
McGovern, T. H.1995.
FARMPACT 5.0. An Economic Simulation
Model for Norse Greenland.
New York: NABO, Hunter College,
City University of New York.
McGovern, T. H., Amorosi, T., Perdikaris, S. and Woollett,
J.
1996. Vertebrate zoo archaeology of Sandnes V51: economic
change at a chieftain's farm in West Greenland.
Arctic
Anthropology
33, 94-121.
McGovern, T. H., Bigelow, G. F., Amorosi, T. and Russell, D.
1988. Northern islands, human error and environmental
degradation: a preliminary model for social and ecological
change in the medieval North Atlantic.
Human Ecology
16,
225-70.
McGovern, T. H., Buckland, P. C., Savory, D., Sveinbjarnard6ttir,
G., Andreason, C. and Skidmore, P. 1983. A study of the faunal
and floral remains from two Norse farms in the Western
Settlement, Greenland.
Arctic Anthropology
20, 93-110.
McGovern, T. H. and Jordan, R. 1982. Settlement and land use
in the Inner Fjords of Godthaab District, West Greenland.
Arctic Anthropology
19, 63-79.
Magnusson,
A.
and Vida lin, P. 1943.
Jaroab6k.
Copenhagen: Hio
Islenzka Frceoafelag.
Magnusson, M. and Palsson, H. 1965.
The Vinland Sagas. The
Norse Discovery of America.
Harmondsworth: Penguin.
Mahler, D. L. D. 1991. Argisbrekka: new evidence of shielings
in the Faroe Islands.
Acta Archaeologica
61, 60-71.
Miller, W. I. 1990.
Bloodtaking and Peacemaking. Feud, Law and
Society in Saga Iceland.
Chicago: University of Chicago Press.
Moore, P. D., Webb, J. A. and Collinson, M. E. 1991.
Pollen
Analysis.
Oxford: Blackwell Scientific Publications.
Nilsson, A. 1943. Den Sentida Bebyggelsen pa Islands Lands-
bygd, pp. 269-310 in Stenberger, M. (ed.),
Forntida Gardar i
Island.
Copenhagen: Ejnar Munksgaard.
Noddle, B. A. 1977. The animal bones from Buckquoy, Orkney,
pp. 174-227 in Ritchie, A., Excavations of Pictish and Viking-
age farmsteads at Buckquoy, Orkney.
Proceedings of the
Society of Antiquaries of Scotland
108, 174-227.
O'Brien, S. R., Mayewski, P., Meeker,
1.
D., Meese, D. A.,
Twickler, M. S. and Whitlow, S. I. 1995. Complexity of

54
T. Amorosi et
ale
Holocene climate as reconstructed from a Greenland ice core.
Science
270, 1962-4.
6lafsson, E. 1991. Islenskt skordyratal.
Fjolrit Natturu!rCE-
oistofnunar 17.
Olafsson, G. 1982.
TorfbCErinn fra Eldaskala til BurstabCEjar.
Reykjavik: pjoominjasafn J[slands.
Rackham, D.
J.,
Allison, E., Colley, S., Lamden, B., Nye, S. and
Owen, M. 1996. Beachview, Birsay: the biological assem-
blage, pp. 161-71 and 186-91 in Morris, C. D. (ed.),
The Birsay
Bay Project,
2:
Sites in Birsay Village and on the Brough ofBirsay.
Durham: University of Durham.
Rowley-Conwy, P. 1983. The animal and bird bones, in Hedges,
J.,
Trial excavations on Pictish and Viking settlements at
Saevar Howe, Birsay, Orkney.
Glasgow Archaeological Journal
10, 109-11.
Sadler,
J.
P. 1987.
The Analysis of Insect Remains from Norse Sites
in the Former Western Settlement of Greenland.
Unpublished
M.Sc. thesis, University of Birmingham.
Sadler,
J.
P. 1991a. Beetles, boats and biogeography.
Acta
Archaeologica
61, 199-211.
Sadler,
J.
P. 1991b.
Archaeological and Palaeoecological Implications
of Palaeoentomological Studies in Orkney and Iceland.
Un-
published Ph.D. thesis, University of Sheffield.
Sadler,
J.
P. and Dugmore, A.
J.
1995. Habitat distribution of
terrestrial Coleoptera in Iceland as indicated by numerical
analysis.
Journal of Biogeography
22, 141-8.
Scott, W. and Palmer, R. 1987.
The Flowering Plants and Ferns of
the Shetland Islands.
Lerwick: The Shetland Times.
Simpson, 1. A., Scott, 1., Kirkpatrick, A. H., Gill,
J.
P. and
MacDonald, A.
J.
in press. Sheep grazing on moorlands of
the Northern Isles of Scotland and its implications for nature
conservation.
Journal of Environmental Management.
Skidmore, P. 1996.
A Dipterological Perspective on the Holocene
History of the North Atlantic Area.
Unpublished Ph.D. thesis,
University of Sheffield.
Snresd6ttir, M. 1991. St6raborg - an Icelandic farm mound.
Acta
Archaeologica
61, 116-9.
Stefansson, S. 1948.
Fl6ra islands
(3rd. edition). Akureyri: Hio
Islenzka Natturufrreoifelag.
Sveinbjarnard6ttir, G. 1991. Shielings in Iceland. An archaeo-
logical and historical study.
Acta Archaeologica
61, 73-93.
Sveinbjarnard6ttir, G. 1992.
Farm Abandonment in Medieval and
Post-Medieval Iceland: an Interdisciplinary Study.
Oxford:
Oxbow Books.
Vasey, D. E. 1996. Population, agriculture, and famine: Iceland,
1784-1785.
Human Ecology
19, 323-50.
Williamson, K. 1948.
The Atlantic Islands. A study of the Faroese
Life and Scene.
London: Collins.
Wylie,
J.
1987.
The Faroe Islands: Interpretations of History.
Lexington: University of Kentucky Press.