‘Wild’ horses in medieval and early modern landscapes of Europe

ABSTRACT Historical mentions of herds of ‘wild’ horses in Europe are frequently taken unquestioningly as a reference for natural grazing in nature conservation. It is often assumed that such herds roamed the wilderness freely and functioned in a socially and ecologically unrestrained manner, uninfluenced by humans. On the basis of a survey of historical (primary) sources, it is examined if these assumptions are justified, or whether these herds should be considered as cultural artifacts. Truly wild horses seem to have disappeared from the North-West European landscape shortly after the Early Mesolithic. Meticulous examination of mentions of ‘wild horses’ in a wide variety of sources from the Early Middle Ages onwards testifies to the fact that these animals had owners and were actually domestic. They were intensively managed in enclosures in the wilderness and deployed in agriculture and warfare. The social functioning of these ‘wild horse’ herds has been influenced for centuries by restricting allowance of stallions, early withdrawal of foals, and protection against predators. In turn, this will have impacted the way these animals influenced the landscape. It does not, however, mean that grazing by free-roaming horses is a bad idea from landscape and nature conservation perspectives. On the contrary, it offers an excellent opportunity to combine protection of both nature and cultural heritage by reinstating the medieval agricultural technique that horse-keeping in the free factually is.


INTRODUCTION
Free-roaming grazers are increasingly deployed in rewilding projects throughout Europe aiming at reintroducing natural grazing (sometimes naturalistic grazing (Hodder et al. 2005, p. 124) as a pivotal ecological process (Linnartz & Meissner 2014, p. 9).The adjective natural to grazing is explicitly used to emphasise the alleged non-cultural character of this measure.The idea behind natural grazing is that it is an inextricable process of 'pristine' ecosystem functioning as opposed to the cultural management technique of seasonal grazing by domestic livestock.Vera (2000) has been an important advocate of the idea of natural grazing.He argues that the original-natural landscape of Europe was half-open and park-like rather than a closed canopy forest.Large, free-living herbivores are believed to be key-drivers for the hypothesised landscape history open natural landscape of post-glacial Europe, on which Vera bases his wood-pasture hypothesis with cyclical vegetation turnover processes.In the past decades, Vera's postulate has led to cutting edge discussions over the question of whether Early-Holocene Europe was indeed a predominantly half-open landscape kept open by grazing or not (see, e.g., Vera 2000, Vera et al. 2006vs. Svenning 2002;Birks 2005;Mitchell 2005;Whitehouse & Smith 2010).Another intriguing question in this context is whether grazing by specific species of ungulates, especially horses and cattle, is really a natural phenomenon or actually an artifact in itself?
In rewilding projects, primitive cattle breeds (Bos taurus; especially Heck, Galloway and Highland cattle) are often used as replacements for the iconic aurochs (Bos primigenius) that became globally extinct in 1627 (Van Vuure 2005).Together with the aurochs, wild horses are considered to be the single most important original grazers in the Holocene landscapes of Europe.Historical sources from the Early Middle Ages to Modern Times regularly mention 'herds of wild horses' that would then still have roamed large parts of Europe and were an integral part of the processes that shaped the landscape.Today, Konik horses and a wide range of other horse breeds are replacing these original 'wild' horses in rewilding projects across Europe (Linnartz & Meissner 2014, p. 9).The herds of 'wild' horses from historical sources, however, are still taken as examples of how natural grazing must have worked in landscapes of the past and serve as historical references for natural grazing processes in naturally functioning ecosystems.In doing so, it is assumed that these 'wild' horses were able to exhibit natural behaviour, devoid of human influence, and were completely free in their activities and behaviour, because this is considered essential in the context of restoring natural processes to the landscape (ibid., p. 22).This article aims to clarify the question whether the assumption of naturally functioning herds of wild horses in historical times is justified or whether such herds were in fact subject to relatively intensive human management and spatial and social constraints, and should consequently be considered as artifacts in the landscape.In other words, is the reference that rewilders have in mind a naturally functioning ecosystem or is it a historical cultural landscape?

MATERIALS AND METHODS
An extensive search for mentions of 'wild horses' was performed in primary sources from the Early Middle Ages to Late Modern Times as well as in secondary, historiographical, historical-ecological, and zooarchaeological literature.The geographic scope was loosely North-Western and Central Europe, chiefly because historical sources for this region were relatively easily accessible.In the pursuit of relevant data and information from (digital) sources, search strings were used such as 'wild', 'untamed', 'forest' and 'roaming' in combination with 'horse(s)', 'mare(s) ' and 'stallion(s)' in different languages (Latin, French, English, German, Dutch) and variants, and, if applicable, in different grammatical cases.The search was extended with historical vernacular spellings of these search strings, like 'wylde pferde' (Old German) or 'wilden perden' (Old Dutch) derived from etymological dictionaries.
Surprisingly few publications in peer-reviewed scientific journals from the twentieth and twentyfirst centuries appeared to deal with this issue; the subject apparently received little attention in natural science literature, including historical-ecological, and restoration ecology journals.Historical documents however, like manor accounts, transfer acts, acts of donation, loan registers, wills, and reports on legal disputes proved to be extremely rich in references of 'wild' horses.Geographically explicit or traceable data from primary and historiographical sources were collected in an Excel-database (SI).This database contains as detailed as possible data about place and time of 'wild' horse mentions, as well as about the consulted sources and relevant quotes.Furthermore, data about types of horses mentioned, as well as relevant contextual information surrounding these mentions, e.g. about grooming, were collected.Separate sheets in the database have been created on relevant data extracted from Domesday Book, given the great importance of this source for this research, and for toponyms related to 'horses in the landscape', derived from place-name etymological dictionaries (especially Gysseling 1960;Berger 1999;Van Berkel & Samplonius 2006;Mills 2012) supplemented by cases that were found by less systematic searches.

EXTANT WILD HORSES?
Recently, Gaunitz et al. (2018) concluded that Przewalski's horses (Equus ferus przewalski) should not be considered truly wild, but rather as feral descendants of domesticated horses herded in Botai in Central Asia.Therewith, the presently widely shared feeling is that wild horses are effectively extinct worldwide, as the tarpan or Eurasian wild horse (E.ferus ferus) is also reported to have become extinct in the wild in Ukraine in 1851, followed by the last specimen dying in captivity in 1919 (Olsen 2006, p. 246;Linnartz & Meissner 2014, p. 13).Koniks are considered at best to be crossbreeds of Polish farm horses and the tarpan, but not purebred wild horses (Van Vuure 2013;Linnartz & Meissner 2014, p. 27).On the other hand, some authors claim that Exmoor ponies should be considered direct descendants or even living representatives of the tarpan (Hovens & Rijkers 2013).In rewilding projects in North-Western Europe, Exmoor ponies would therefore be amongst the most suitable breeds for natural grazing (Hovens & Rijkers 2013;Linnartz & Meissner 2014, pp. 24-30).Evidence for the claim that Exmoor ponies descend directly from the tarpan is said to be derived from a broad spectrum of studies including research on metatarsal bones, size and form of molars and jaws, blood proteins, DNA, and historical sources.It is argued that Exmoor ponies were influenced effectively by humans no earlier than approximately 1900, from which moment on it was determined which stallions were or were not allowed to cover the mares; earlier influence of humans on the breed is considered to be minimal.14 C-analyses of zooarchaeological remains of wild horses in Europe (Sommer et al. 2011) show that there are no finds of wild horse in England younger than c. 10,500 years old, although Bendrey (2012) mentions a single radiocarbon date that suggests wild horse populations may still have been present in England somewhere between c. 9000 b.c. and 2000 b.c.In general, however, zooarchaeology provides no proof that wild horses still roamed England after the Early Mesolithic.This is in concordance with findings throughout continental Europe where wild horses also disappear from the zooarchaeological spectrum during the Mesolithic (Sommer et al. 2011).During the Holocene there were expansions in the distribution of wild horses in Europe (following contractions) but England could not be recolonised after approximately 8000 b.c.since Doggerland no longer formed a connecting land bridge with continental Europe (Gaffney et al. 2009).The increasing number of horse remains in English archaeological sites from the Bronze Age onwards is exclusively attributed to domestic horses.Distinguishing between wild and domesticated horses on the basis of archaeological findings, however, is surrounded with uncertainties (Bendrey 2012).The question therefore remains whether (small) populations of truly wild horses could have survived in England unnoticed for thousands of years without leaving traces in the subfossil records.Neville (2006, p. 139) does not rule out that there may have been still some genuine wild populations during the Anglo-Saxon period; Hovens and Rijkers (2013)   well as numbers of cattle and sheep, watermills and salterns.Next to horses kept at the estate, there are also references to equae silvestrae (also silvaticae and silvestres; Darby 1977, p. 165; all discussed here under the joint heading of silvestrae) as well as to equae indomitae, usually translated as forest/woodland mares and unbroken/untamed mares, respectively.In total, Domesday Book holds ninety-eight mentions of such horses (Fig. 1), of which forty-four are located in the shires of Devon and Somerset, the heart of present-day Exmoor National Park and home to the Exmoor pony.The equae indomitae appear to be restricted to south-west England while equae silvestrae are also reported from the east, notably Norfolk and Suffolk.Equae silvestrae seem to refer to horses that lived in woodlands or, if one prefers, in the wood-pasture sensu Vera (2000).Merely from the name equae silvestrae, however, one cannot conclude that this actually concerned wild or even feral horses; it only means that these horses lived or foraged in woodlands.Likewise, one cannot interpret equae indomitae as being wild horses by definition, only as horses that were not made used to the bridle and trained.Indomitae may even have meant unaccustomed or unsettled, referring to the fact that they were not used to living in the manor.But also such horses can be subjected to selection and intensive care.To judge whether these equae concern truly wild horses or not, one must realise that Domesday Book was primarily intended as a basis of taxation, which means that these horses must have had owners who could be taxed.Nowhere in Domesday Book are there references to other large wild ungulates like red deer or wild boar (Williams & Martin 1992), simply because such animals were not considered to have owners liable for taxation.It is therefore extremely unlikely that the horses in Domesday Book can be considered wild animals.Against the assumption that equae silvestrae/indomitae were really wild horses also speaks that in (almost) all cases in Domesday Book exact numbers of these horses are mentioned.It would have been virtually impossible to establish this for truly wild horses.Owners, however, would have known quite well how many horses roamed their estates at any one time.
There is also other proof that 'wild' horses in medieval England had holders.In a.d.950 the Anglo-Saxon lady Wynflaed left a personal bequest to a beneficiary of 'hyre dael Þera wildera horsa Þe mid Eadmaeres synt' ('her share of the wild horses which are with Eadmer's') (Whitelock 1930, pp. 14-15;Hooke forthcoming).From this it can also be inferred that 'wild' horses had some function, at the very least as status symbols.In that context they were also regularly part of gifts and legacies from the seventh to the tenth century throughout north-western and central Europe (Whitelock 1930, p. 51;Davis 1987, pp. 76-7;Franz 2006, p. 45).It seems highly unlikely, however, that status symbol was the only or even prime function of these horses, and that they had no direct practical use but were subject to taxation at the same time.This, of course, raises the question of which other functions these horses may have fulfilled and why they were allowed to roam freely in the wilderness instead of being kept stabled.
Since eating horse meat was prohibited in a.d.732 by Pope Gregorius III, it seems highly improbable that horses were kept in the wilderness to be hunted.There is also little zooarchaeological evidence that medieval horses were slaughtered for consumption.On the other hand, it must be mentioned that in medieval and early modern Prussia, possibly truly wild horses were still hunted for hides and meat.According to Rünger (1925, p. 223), this was also the case in twelfth-century Denmark.And from a list of graces of the Abbey of St Gallen, Switzerland, from around a.d.1000 comes the quote: 'sit feralis equi caro dulcis in hae cruce Christi' ('This sweet wild horse flesh in the cross of Christ'), which shows that at the time of Domesday Book horse meat was still eaten, even by monks, despite the Pope's ban.Therefore, also in medieval England occasional hunting of horses cannot be ruled out completely.
The most logical potential function of horses, however, is that of draught animal for agriculture, even though horses were not used for ploughing in Anglo-Saxon England (Keefer 1996, p. 118).The shoulder-collar, necessary for harnessing horses to the plough, came into use in Europe from about a.d.800 but did not become widespread before the eleventh or even twelfth century (Slicher van Bath 1963, p. 63;Le Goff 1988, p. 213;Keefer 1996, pp. 119-20).Until then, ploughing was performed with teams of oxen that were yoked to the plough, of which Domesday Book holds many references (Darby 1977, pp. 122-6;Williams & Martins 1992).Horses were used, though, for harrowing and other light agricultural work, but also for driving horse mills, riding, pulling carts, and as pack animals (Williams & Martins 1992;Keefer 1996, p. 118;Hall 2005, p. 25).In hilly terrains, pack horses were virtually the only mode of transport for large quantities of goods (Le Goff 1988, p. 213 Gerhold 1993).
In twelfth-to late fifteenth-century accounts of English manors, lesser horses are referred to as affri and stotti (Thorold Rogers 1866-1902).Other categories of horses mentioned in manorial accounts are destriers, rounceys, palfreys, and hackneys that must have been, given their mentioned values, breeds of horses far superior to the affri or stotti (see also Hooke 2015, p. 263).These more expensive horses were kept stabled at the manor (Neville 2006, p. 138;Le Goff 1988, pp. 122-4) since they required intensive daily care and training to keep them in perfect shape for agriculture, transport or battle.When going to war, for each destrier at least three additional horses were needed: two riding horses for the knight and his squire, who led the destrier, and at least one for carrying or pulling carts with armour and supplies (Rünger 1925, pp. 229-32;Davis 1987, p. 79).On large campaigns, such as the Crusades, the required numbers of horses must even have been much larger to meet the daily needs during the long journey.The significance of large numbers of horses for warfare cannot be underestimated: the strength of the medieval armies depended chiefly on the numbers of their mounted horsemen and the number of transport animals that could be raised to support them (Clutton-Brock 1992, pp. 138-9;Hall 2005, p. 24;Williams 2015, p. 199).
As large numbers of horses were only needed at peak times in the agricultural season, or during periods of war, it was not necessary for all horses to be available on a daily basis.Unlike cattle, horses require intensive care and exercise when kept in stables.They need to be fed and watered at least two or three times a day.In addition, they require feed in the form of oats which were expensive and not always easy to obtain (Le Goff 1988, p. 213).Horses that were not actually used for most of the year therefore would have required expensive care without any direct benefit.Free-roaming horses, on the other hand, can take care of themselves perfectly well and do not need further attention if they have enough space (Zimmermann 1999).It may have been such horses that were left to roam the wilderness and fend for themselves, that formed the herds of equae silvestrae and equae indomitae mentioned in Domesday Book.They were only taken out of the herd when needed and were broken in (if necessary) and put to work or sold as affri or stotti.All equae, be they indomitae or silvestrae, almost certainly also served as broodmares for the production of large numbers of workhorses at the lowest possible costs (Franz 2006, p. 43).Breeding may even have been their primary function.They were thus not prestige objects in themselves, but the possession of these animals was of great importance to the British medieval elites.

FREE-ROAMING HORSES IN EUROPE
Referring to and discussing equae silvestrae/indomitae almost exclusively within the context of Domesday Book may suggest that such horse-keeping practices occurred only or predominantly in medieval England.Data from other sources than Domesday Book make clear that this was certainly not the case.Figure 2 shows the distribution of mentions of 'wild' horses (in diverse phrasings) throughout Europe from the Early Middle Ages to Early Modern Times.References concern either direct or indirect mentions.An example of a direct mention comes from a document from 1288, according to which German noblemen promise that their 'wild' horses ('equos nostros silvestres'), which roamed the Kliedbruch marsh area, would be mingled with those of the Abbey of Meer (Venner 1985, pp. 287-8).Indirect mentions concern, for instance, references to feudal rights to graze and breed 'wild horses', an example of which comes from the loan register of the Count of Guelders (the Netherlands) of 1326 which mentions a certain Jacob of Myrlaer who held in fief: 'den ganc van den wilden perden in Echterwalt' ('the right to let wild horses graze in the forest of Echt': Venner 1985, p. 285).The various sources show that 'wild' horses were widely spread in Europe, from Sweden to Spain, and from Scotland to Ukraine.In most cases they were referred to as equae indomitae, followed by equae silvestrae, but sometimes they were also known as equi vagi and equi silvatici, or by local vernacular names.Pinters von der Au (1688, pp.2-3) reports of 'wilde Pferde' ('wild horses') from Hungary, Germany and Poland, among others, but also mentions that these horses should not be considered truly wild ('nicht wilde Pferd zu nennen') because they were withdrawn from the herd at a certain age, broken in and put to work.
From the sources, it can also be inferred that in medieval times herds of free-roaming horses were often found in riverine areas or in (associated) swamps and marshes (Bruch in German, broek in Dutch, related to the English brook and derived from the Germanic broka) and in other areas unsuitable as arable land.From Eastern Europe there are references to semi-wild horses in medieval Bohemia, Moravia and Poland along the rivers Elbe and Vltava (Jireček 1864, p. 43;Šmelhaus 1980, pp. 39-41) and Vistula (Szymczak 2018, p. 233), respectively.Nearly all medieval sources reporting on the occurrence of free-roaming horses in the Netherlands and the western part of Germany relate to the area directly east of the rivers Meuse and IJssel, the Netherlands, and along the River Rhine and its tributaries in Westphalia, Germany.Davis (1987, pp. 74-5) briefly elaborates on an example of medieval horse breeding in the limestone-rich landscape of the River Seine in Normandy, France.
In their natural state river floodplains and marshland are less suitable for arable agriculture because of high groundwater tables and flooding risks.Historically, high groundwater tables also prevented succession to closed forests, as did natural river dynamics which regularly set back succession, resulting in a half-open landscape suitable to support free-ranging herds.Finally, riverine and particular marshland areas are also favoured by horses because of the availability of calcareous grass (Davis 1987, p. 74;Whitehouse & Smith 2010).

HISTORICAL MANAGEMENT OF 'WILD' HORSE HERDS
Direct and indirect historical mentions of freeroaming herds make clear that they must have represented an important economic value.How can this be reconciled with the fact that, in rewilding literature, historically mentioned herds of horses are often described -implicitly or explicitly -as having a high degree of freedom in their comings, goings and doings?Would that not have meant taking unnecessary risks for the horses and therewith for their holders?Did the horses' owners really put not much effort and means into the management of the herds?To what extent did the freedom of herds actually extend, both spatially and in terms of social functioning?Below, medieval and early modern practices of keeping free-roaming horses are elaborated, with special reference to the role of custodians, the existence and functions of enclosures, the regulation and management of stallions, and the way of managing and harvesting 'wild' horses.custodes Domesday Book mentions a servant entrusted with the care of the King's forest mares in Kingston upon Thames.Although it does not become apparent what this care actually included, it does signify that these forest mares were not roaming through the wilderness without any human supervision.Several historical sources from continental Europe from the twelfth to fifteenth century also mention custodes (caretakers) of free-roaming herds (e.g., Czech Republic, Germany and Hungary).In 1294-1295, for example, in Zieuwent, the Netherlands, an amount of rye was given servo costodienti equos silvestres ('to the servant taking care of the wild horses').Other examples are the mentions in accounts of a custos vagorum equorum ('groom of roaming horses') in the Dutch Nederrijkswald in a.d.1340 and of the honorary duorum custodum equorum indomitorum regis ('of two caretakers of the King's unbroken horses') in Aberdeenshire, Scotland in a.d.1438 (see also SI for more examples).Jireček (1864, p. 43) speaks in general terms of custodes equorum during the Middle Ages in the border regions of Bohemia, Moravia and Silesia; Šmelhaus (1980, p. 41) refers to Bohemian horse keepers as agazones.
The monasteries in particular were involved not only in receiving 'wild' horses as gifts or legacies, but also in buying and selling horses (Davis 1987, p.78).In the High Middle Ages they also became central players in the breeding of horses.Interesting from this perspective is the mention by Dudík (1865) of the Strahov monastery near Prague which received in 1143 'silvam, ibidem equas, quas dicimus ammissarias' ('a forest, in which horses, which we call broodmares').Often located in the same remote wilderness where the horses roamed, monasteries appeared ideal for supervising the herds to breed and sell the horses, with monks grooming them (Davis 1987, p. 78).
areas for semi-wild horse keeping The Middle Low German word Stôt, clearly related to the English stud, the Germans Gestüt and the Dutch stoeterij, is often mentioned within the context of feudal breeding rights of horses in the Middle Ages.Originally, the word was interpreted as a fenced area for horses in the free ('Umzäunung für Pferde im Freien, Pferdebahn'), indicating that the spatial freedom of the free-ranging horses was certainly not unlimited.Sometimes, tautologies of 'Stôt' can be recognised in toponyms, like in Stuttgart (Germany) or Stolegarda (Italy); both names mean 'enclosure for a stud' (see also below and SI for more stud-related toponyms).Davis (1987, p. 70) mentions fenced areas for horse keeping also for France and states that these enclosures were primarily meant to keep the mares from unwanted external influences.Later, the same word Stôt was also used to refer to the horses themselves ('in dieser Umzäunung behaltenen wilde Pferde'); cf. also the earlier mentioned stotti ('horses withdrawn from the stôt').
'Wild' horses were fenced in because unrestricted spatial freedom would have had four major disadvantages.First, it would have been much more difficult to round up horses when there was need for them; secondly, the owners would have had no control over reproduction; thirdly, herds would have been exposed to predation and competition, the latter either with large wild ungulates, or with other herds of horses or cattle; and fourthly the horses would do great damage to arable crops.It is therefore an illusion to think that the horses would not have been confined within some kind of enclosure.According to Davis (1987, p. 70), keeping the enclosures in first-class repair was one of the most important management measures for the herds.The enclosures were probably fenced by palings, earthen banks or stone walls, and/or by hedges of thorny shrubs such as hawthorn, and provided with a ditch (Cantor & Hatherly 1979, p. 72;Wager 2017, p. 178).
In general, Domesday Book does not mention that equae silvestrae/indomitae were kept in fenced areas.It does, however, mention the words parcus (often with the addition bestiarum silvaticarum ('for beasts of the forest') or ferarum silvaticarum ('for wild animals of the forest') and haia, both being usually interpreted as enclosures for the hunt (Darby 1977, pp. 201-6).The word parcus is the Latin word for park; the word haia (latinised form of Old-English haga, but see also Hooke 1989, pp. 123-6;Wager 2017, pp. 169-71) equals the present-day English place-name element hay and is interpreted by Darby (1977, p. 204) as a 'hedged enclosure constructed to control the driving and capture of animals'.In only a few cases specifications of haiae are given; on folio 260recto of Domesday Book, for instance, capreolis capiendis ('for the capture of roe-deer ' Darby 1977, p. 204).In most cases, however, it remains obscure for precisely which purposes the parci and haiae were used.Although it is often assumed that their main function was hunting ground, there are also scholars who are of the opinion that this may have been a side or even secondary function (Wager 2017, pp. 178-9, pp. 191-3).From our perspective it is interesting that the parci and haiae mentioned in Domesday Book mostly do not coincide geographically with the areas from which free-roaming horses are reported (Fig. 3 and SI, tab Domesday overview).The equae silvestrae/indomitae are chiefly mentioned for the Norfolk and Suffolk shires in the east, and Cornwall, Devon and Somerset in the southwest of England; parci are mainly reported from Hampshire, Hertfordshire, Worcestershire, and Sussex; the haiae from Herefordshire, Shropshire and Cheshire.Parci and haiae have been interpreted to be local names for the same type of enclosures and both terms are often used interchangeably (although there may have been subtle differences; see Wager 2017, pp. 183, 191).References to equae silvestrae and indomitae especially relate to horses kept within enclosures, which could actually be parci and haiae as well.Such inconsistencies in Domesday Book may be due to the fact that surveys in different parts of England were performed by different investigators (inquisitores) (Darby 1977, pp. 5-6).Indications that parci and haiae may have been used for keeping horses comes from old place-names like Horsehay in Shropshire, England (for more toponyms, see below).
The Liber Exoniensis, a satellite text of Domesday Book, gives a single indication of the function of a parcus, namely at Donyattin, Somerset (fo.92verso).This parcus is said to have comprised '1 cob; 12 unbroken mares; 9 cattle; 70 sheep; 30 goats', making this the only Domesday parcus for which there is evidence that 'wild' horses were kept in such enclosures.It also makes clear that within the parci, besides unbroken mares, other domestic animals were kept.Also in the seventeenth century, claims on grazing rights in the New Forest 'horse beasts' are mentioned, next to 'rother beasts' (bovines) and other domestic livestock (Anon 1853).Apparently 'wild' horses were bracketed together with cows and pigs.In only one case, the measures of a haia are given in Domesday Book (Donnelie, Warwickshire; fo.240recto): 'half a league long and half a league wide', giving a total acreage of approximately 600 ha, roughly matching the minimum required surface area to support a self-sustaining herd of free-roaming horses (Renes 1999, p. 184;Linnartz & Meissner 2014, pp. 33-5).Some parci and haiae may thus have served primarily as hunting grounds but, at the same time, also as fenced areas for domestic stock keeping and breeding, including free-roaming horses, for the whole or a part of the year.Also elsewhere hunting and horse breeding was combined, like in the Merlenbroek in Dutch Limburg, where -according to the report of a dispute about a fief in 1456 -there is mention of a combined 'peerdenstoet ende wiltbaen' ('horse stud and confined hunting ground') (Venner 1985, p. 286).In other parci and haiae, horse keeping may even have been the main function, like the derhage at Ongar, England (Sykes 2007, p. 60) of which is written: 'Þat derhage 7 Þat stod Þe ic Þer habbe' ('the derhage and the stud which I have there' (Whitelock 1930, pp. 82-3).'Der' in derhage is in this context often misleadingly translated as 'deer' and may well be the reason why these landscape elements have often been interpreted as hunting facilities for roe deer and red deer, but it actually has a much broader meaning, namely 'animal', cf.German 'Tier' and Dutch 'dier' (see also below under 'Horse' toponyms).If enclosures like haiae and parci were in some cases indeed used for horse keeping, be it as primary or as secondary function, 'wild' horses may have been present in far more regions of medieval England than can be inferred from Domesday Book directly.regulation of stallions According to Van der Sijs (2010) the Old High German word stuot means 'half wild herd of horses, consisting of one stallion and several mares' (cf.Middle Low German Stôt in the previous section).Apparently, stuot does not refer to a natural herd consisting of both harems and groups of stallions which do not (yet) participate in reproduction, but only to the former.This is also in concordance with the fact that Domesday Book and many other sources speak about 'wild' horses in the feminine form equae, 'mares' (Darby 1977, p. 165;Williams & Martins 1992) and not of the more neutral equi (as wrongly suggested by Hovens & Rijkers 2013, p. 130).This indicates that at least the number of stallions was regulated in the free-roaming herds.Neville (2006, p. 139) also states that stallions were not allowed to run freely with the mares and that most of the Anglo-Saxon 'wild' horses were probably not entirely 'unimproved' by human selection and interference in their breeding.Keefer (1996, p. 123) mentions eleventh-century wills which contain evidence of well-established stud-farms and suggests the existence of a substantial breeding programme in operation by the time of the Norman Conquest of England.The Carolingian agricultural regulations of Francia, known as the Capitulare de villis, stipulated that old and sick stallions were not allowed to cover the free-roaming mares, and stallions were carefully selected.It is well known in horse breeding that regression rapidly occurs in certain meticulously selected characteristics, especially size, if it is not precisely determined which animals are and which are not allowed to take part in reproduction (Davis 1987, pp. 69-70).
In the twelfth/thirteenth-century German epic poem das Nibelungenlied, the hero Siegfried hunts a wisent, an elk, four aurochs and a 'shaelch' in the Germanic forest.The latter is interpreted as the old leading stallion of a herd of mares; cf.German: Beschäler 'stallion' (Franz 1967).These old leading stallions are also known under the names 'schelros', 'scheleros', 'schele', 'beschel', 'schelhengst', and 'schaelgaewl' (Rünger 1925, p. 214).The word can also be found in ninth-and tenth-century Dutch charters in which the hunt for the 'scelo' is reserved for the church of Utrecht.According to a kitchen account from 1537 (district of Lippe, Germany) a shot 'Schelch' even ended up in the cooking pan (Krüger 1944, p. 216), thus violating the church rules on consumption of horse meat.Assuming that by that time truly wild horses were indeed extinct in Western Europe, this may have concerned writtenoff, and subsequently released (or escaped) stallions, roaming the wilderness to be hunted by members of the higher societal classes.In some cases, however, it cannot be ruled out that it concerned truly wild (and thus not feral) stallions, especially in North-Eastern Europe.
According to Davis (1987, p. 70) keeping unwanted stallions away from the mares was one of the main reasons why the enclosures were surrounded by stallion-proof fences in the form of palings and/or thorn hedges.During Charlemagne's time stallions were kept apart and put in with the mares only during the mating season.The same carefully selected, high-quality stallion could thus serve several groups of fenced mares, one after the other (ibid., p. 72).The selected stallions were probably also not always unbroken.Indirect evidence for this comes from the will and testament of a certain Wulfric, an Anglo-Saxon chieftain who around the year a.d.1000 bequeathed to the monastery of Burton 'an hund wildra horsa and sextene tame hencgestas' ('one hundred wild horses/ mares and sixteen broken stallions' : Whitelock 1930, pp. 50-1).
There are also other reasons that stallions would most certainly not have been kept within the same fenced area as the mares.A surplus of stallions would lead to mutual fights between them and to social upheaval in the herds.In Salian Law (Lex salica) during the reign of the Merovingian ruler Clovis (beginning of the sixth century) the number of mares per stallion was determined to be from seven to maximum twelve (Franz 2006, p. 43); from other parts of Europe we read of ten to fifteen (Davis 1987, p. 70) to a maximum of twenty to thirty mares and their foals per stallion (Renes 1999, p. 184).Apparently, the number of mares per stallion was strictly regulated.managing and harvesting the herds Since herds consisted primarily of a limited number of mares (the average number mentioned in Domesday Book is fifteen) and were living in restricted fenced areas, spontaneous exchange did not occur and they may have suffered from inbreeding in the long run.To bring in new blood, now and then animals from elsewhere were brought into the herd, witnessed, for example, by Boecke (1536, p. 41) mentioning the mixing of domesticated and 'wild' horses in Scotland and by the 1288 document mentioned before, according to which the Lords of Meurs mingled their 'wild' horses with those of the Abbey of Meer (Venner 1985, p. 288).
Also, the owners or custodians will otherwise certainly have interfered with the social functioning of the herds.Davis (1987, p. 72) describes in detail the tailored procedures according to which the herds were managed and how this offered opportunities to increase the 'yield' (called selvagium under the Merovingians : Franz 2006, p. 43).Herds which experience no competition or predation can spend much more time on grazing and are in a much better shape than herds which constantly have to be on the qui vive.This was an additional reason to restrict herds to a fixed area to keep competitors and predators at bay.
Under natural circumstances colts are driven out of the herd by the leading stallion when they are about two years old.Many of them subsequently die before they reach maturity.By removing them at an early age and placing them in other enclosures where they were fed grain and hay, they had a much better chance of surviving the winter.References to semiwild herds of foals go back to 1555 when Olaus Magnus mentions them roaming the wood-pastures in Sweden until their third year, from which age onwards they were trained for work (Foote 1998, p. 858).After removal of their foals, the fertility rate of mares increases markedly.They no longer have to nurse them and can retain all nourishment of their grazing for themselves.Additional feeding also increases winter survival of the adult mares (Davis 1987, pp. 71-2).Interesting in this context is that Jireček (1864, p. 43) makes mention of nutritores equorum ('feeders of horses') when referring to custodes in twelfth-century Bohemia and Moravia.
Foals withdrawn from the herd could be broken and trained as riding, pack or draught horses for their owners (Franz 2006, p. 44).Colts were then mostly also castrated (Rünger 1925, p. 216) while fillies could of course also serve as unbroken broodmares by adding them to the dam herd or by starting a new breeding herd (Davis 1987, p. 72).Young horses were also sold, witnessing a 1340 account mentioning the proceeds of the sale of seventeen foals from the herd of wild horses of the Duke of Guelders, the Netherlands (Nijhoff 1830, p. 19).Profits from the sale of equorum indomitorum are also reported in Scotland's fifteenth-century royal accounts (e.g.Burnett 1884).These accounts also make clear that men were paid to round up the horses and drive them to the market where they were sold.Wild bred horses were even one of the three main export items of Bohemian merchants in the ninth century (Šmelhaus 1980, pp. 39-40).
According to both Davis (1987, p. 72) and Franz (2006, p. 46), many of the management measures and practices described above had already been adopted in medieval Europe by the eighth century.In the Capitulare de villis (chapters 13 and 14) some of these measures were translated into regulations for the stewards of the royal estates.In conclusion, management of herds of 'wild' horses (separating mares and stallions, selection of breeding stallions, removing foals at a young ages, and supplementary feeding) was common practice in Europe from the early Middle Ages onwards and aimed at increasing reproduction yields and the economic profits of horse breeding in the free.Therefore, it cannot be claimed that these herds were spatially and socially unrestricted in their goings and doings.

'HORSE' TOPONYMS
The form, size and function of horse breeding in the wild varied greatly over time and place, and different breeding conditions will have resulted in equally different horse qualities.In the Capitulare de villis, for instance, breeding conditions were strictly regulated, providing high-quality horses.We see the same in the breeding farms of the German Order in thirteenth-sixteenth-century Prussia.Here, separate studs for high-quality military horses and agricultural workhorses were operated.Military horses were bred and reared in carefully managed studs ('grohse (large) Stut') while workhorses were extracted from the 'cleyne (small) Stuts', also called 'wilden Stut' or 'Ackergestüt' (Rünger 1925, pp. 256-7).This may also have been the case for the sometimes pairwise occurring estate enclosures in Sweden named 'Stora Diuregården' (great 'animal gard') and 'Lilla Diuregården' (little 'animal gard') that go back to at least the early seventeenth century (Ahrland 2011, p. 76).Thus, even at a short distance from each other and during the same time period, studs varied in method and intensity of breeding.From descriptions in the historical sources of studs in the free as well as of herds of 'wild' horses in the form of equae indomitae or equae silvestrae, it can mostly not be derived to what extent horse breeding was actually regulated.However, given the great need for common riding and workhorses, the less controlled 'small studs' in medieval and early modern Europe will have been far in the majority.The landscape will therefore have been influenced on a large scale by these hundreds of thousands of horses roaming more or less freely through the countryside.We will probably never know the geographical extent of this because the availability of sources is limited in this sense.Toponyms, however, can help us gain some additional insights.It should be noted, however, that in the case of toponyms, it is often even less clear whether we are dealing with completely free-roaming horses or with more regulated conditions.
Before delving deeper into horse-related placenames, an issue must be resolved that is of great importance to correctly assess the toponyms discussed here.Many horse-related place-names are so-called complexes in two parts, the first of which refers to the animals themselves and the second to the geographical entity.An example is the English place-name Horsehay which consists of horse+hay.In many toponomic dictionaries and other literature, place-names beginning with Maar-, Mar-, Meer-or Mer-, be it in England, Germany, the Netherlands or Belgium, are interpreted as toponyms related to derivations of the Proto-Germanic *mari-, meaning 'lake, marsh or water basin' in general.However, this leads to strange meanings of some place-names such as the Dutch place-name Mierlo (allegedly from *mari+*lauhaz) which would mean 'enclosure in dry forest in a lake/ marshland', or the English Meersbrook: 'marshland near the marsh'.A much better explanation for place-names beginning with Maar-, Mar-, Meer-landscape history or Mer-is that the first part refers to derivations of the Proto-Germanic word *marhijō-, 'mare' (Old English meare, mere (Mercian), myre (West Saxon), Old Saxon meriha, Old Norse merr, Dutch merrie, Old High German meriha, German Mähre).Especially the dozens of place-names that connect Maar-and the like with enclosures (-gard, -lauhaz, -hlaeri, -tūn, -wïc) then suddenly make a lot more sense.These would be historical fenced-in places where mares ran.However, because the link between the occurrence of free-roaming herds of mares and place-names has rarely or never been made, these potential meanings have hardly been recognised.I am only aware of a few place-names of which the toponymic sources indicate that they are derived from 'mare' (see SI for references), such as the German Marl (from ninth-century Meronhlara 'enclosure for mares in a marshy forest'), the also German thirteenth-century Merienrode ('clearing for mares') and the eleventh-century French-Flemish Merielant ('land for mares').Surprisingly few place-names were found that connect 'mare' with 'haia'; notably only Meerhage, Belgium, and Meerhay, England.However, there are place-derived surnames of persons that show that there must have been more of such field-or place-names (e.g.Van Meerhage, Van Merhegen, Meirhaeghe).
Many Maar-, Mar-, Meer-or Mer-names are connected with derivations of Proto-Germanic *felthan (-field, -felt, -velt, -veld, etc).Proto-Germanic *felthan is related to 'to fell' and probably has the original meaning of 'forest clearing'.The Slavic equivalent of 'field' is 'pole' (Polish and Czech); the Slavic word for mare is kobyła (Polish) or kobyla (Czech).Combinations of both words form historic Polish place-names like Kobylego Pola (nowadays Kobyla Wieś) and Cobilepole (Kobylepole) and the Czech place-name Kobylí pole (nowadays just Kobylí).In this sense these place-names are unmistakably equivalents of the Germanic 'mare fields' and not of 'marsh fields' or 'lake fields'.A last indication that Maar-, Mar-, Meer-or Mernames do refer to mares and not to swamps or lakes, is the bilingual Polish/German place-name Kobylata/Stuttfeld in which both the Slavic word for 'mare' and the German words for 'stud' and 'field' are recognisable.An interesting finding in this regard by Szymczak (2018, p. 233) is that along the Vistula near Cracow in thirteenth-fourteenthcentury Poland, the owners of the 'mare fields' had the rights to let their horses graze in winter on nearby arable land belonging to other landowners.Bohemia knew similar grazing rights, which were laid down in the jus et libertas campi iumentorum from 1281 (Šmelhaus 1980, pp. 39-41).This privilege applied only to semi-wild broodmares (jumenta, eque indomite, eque emissarie), which had to be branded with their owner's mark.These horses were not allowed to be shod, because then they were considered draught mares by law and lost their privilege.It proves that the horses from the 'mare fields' were actually free-roaming, 'wild' mares, quite similar to the equae indomitae and silvestrae, and also clarifies once more that these horses had owners.
The distribution of horse-related place-names in Europe is given in Fig. 4 (see also SI for more detailed information).A total of 307 records were collected comprising 196 different places.For further analysis not all records were used.Celtic and Slavic names were excluded because their exact meaning could not always be determined, leaving only Germanic names.Places that were obviously horse-related, but from which the geographical meaning was not clear or deemed not relevant, were excluded as well.Of identical attestations from different years, only the oldest was kept.In the end, this yielded 236 historical attestations of 155 different places in eleven countries across Europe.
Of the 155 different place-names, twenty-six referred to 'stud', forty to 'horse' in general, nine to 'stallion' and two to 'foal', meaning that the remaining seventy-eight place-names referred to 'mare', clearly showing the dominance of mares in the landscape (especially since 'horse'-toponyms can also refer to places that were factually dominated by mares).
Forests (in the present meaning of the word) and forest clearings, pastures and enclosures, including -hain, -holt, -wold, -wood, -lēah, and -rode, make up thirty-six of the place-names' geographical second part of the toponomic complexes; heathland/commons only in two cases.Marshland and other wetlands (excluding marshy forest), especially in the form of -brook-, make up for fourteen of the 155 placenames.Fields, either in the form of -kamp ('fenced fields') or in the form of -field proper are present in fifteen place-names.Especially the earlier mentioned combination of mare+field is often attested.
Altogether, this shows that horses were kept in a wide variety of environments, from wet to dry and from wooded to completely open, but mostly in fairly controlled conditions, i.e. within fences.In the modern landscape, not only this scenic and ecological diversity has disappeared, but also most of the herds of 'wild' horses themselves, although they are sometimes brought back as part of ecological restoration.Often, however, only the place-and field-names still recall their former presence (see also Hooke 2015, pp. 262-3).

FREE-RANGING HORSES IN THE RECENT PAST
Evidence for breeding horses in the wild before the Middle Ages is practically non-existent.The closest is the mention of 'common horses' in De Re Rustica, which according to the Roman author ) were grazed everywhere and for which there was no fixed season for breeding.But mentions of free-roaming horses in the Late Modern period are almost equally scarce, especially considering the overwhelming amount of written sources available.But still, free-roaming horses were present across Europe.Le Franq van Berkhey (1769, p. 44) mentions them along the rivers Lek and Maas in the Netherlands.Mares and foals were left to fend for themselves until they were big and strong enough to be sold and put to work.In other parts of eighteenth-century Holland and Germany (e.g.near Vaassen, Freckenhorst and Recklinghausen) horses were kept in the open only in winter (Snelting 1901, pp. 89-90); Youatt (1831, pp. 20-1) reports the existence of semi-wild herds in Sweden, Finland and Iceland.Here animals were taken from the herd when necessary and used for various purposes.Herbert (1863, p. 68) reports similar practices for the Shetlands, and Winkler (1868, pp. 306-7) for the department Brenne in France.Pauly (1828, p. 333) stills mentions the presence of 26,000 horses in Iceland of which 'eine nicht unbedeutende Zahl' ('a not inconsiderable number') roamed freely over the island and could be captured by farmers when they needed horses.However, the appreciation of such horses declined noticeably in the course of the nineteenth century.The best example of this is given by Youatt (1831, p. 42) who mentions that in the woods and common lands around London there were herds of wild horses which he described as 'ragged, wild and mongrel horses' and that 'a good horse is hardly ever bred there by chance'.The causes of the decrease in the number of herds are still a matter of conjecture.Undoubtedly, the need for more arable land was an initial cause, leaving less wilderness land for the horses.Better, more controlled breeding programmes may also have contributed to the abandonment of free breeding.
In the twentieth century, railways and motorised traffic took over the functions of many workhorses.Apparently, the once respectable status and much needed supply of 'wild' horses came to an end and free-roaming herds were eventually reduced to just a few, such as the 'wild' herds of the Camargue in France, the Merfelder [sic!] Bruch in Germany, the Hortobágy National Park in Hungary and, of course, the Exmoor National Park in England.

CONCLUSIONS
Scrutinising available historical sources can contribute to a better understanding of the role of freeroaming horses in cultural landscapes of medieval or even earlier origin, including their significance for ecological landscape-forming processes.In this way, such insights can also shed light on nature and landscape conservation issues, in this case as to how far historical herds of free-roaming horses in Europe can be regarded as truly wild and to what extent they contributed to natural processes in landscape functioning.Although this study concerns large temporal and geographical scales roughly covering a millennium and ranging from Iceland to Spain and from Latvia to Great Britain, the historical insights presented clarify that grazing by free-roaming herds of horses can by no means be considered as purely natural.Zooarchaeological data make clear that wild horses probably became extinct in most of Europe during the Early Mesolithic; only in North-Eastern Europe did the tarpan survive into historical times (Van Vuure 2013).Domestic horses were introduced in Western Europe during the first millennia b.c.(Mallory 1989).Consequently, the presence of herds of free-running horses in Europe during the last two millennia or more is not factually the case despite human presence but is a consequence of human activities and should be considered artifacts in the landscape.These herds have been subject to human influence for many centuries, with regard to both the actual spatial freedom they had and their social functioning, although this influence may have varied considerably across time and space.The surface area that the herds had at their disposal was restricted due to both management requirements and economic objectives.The social functioning of the herds has been deliberately influenced for centuries by temporal allowance of selected stallions, premature withdrawal of colts and fillies, protection of herds from predators and competitors, and additional feeding; all in order to maximie herd proceeds.
In most, if not all, scientific studies underlying rewilding projects, no or only little attention is paid to the cultural-history of free-roaming herds.It is implicitly assumed that herds of 'wild' horses present in medieval Europe behaved in a truly wild manner, although this is not supported by archaeological evidence or by written sources.The reconstruction of the historical-ecological functioning of medieval landscapes given by, for example, Vera (2000) is thus in fact a description of the role that domestic horses and cattle played in the socio-economically driven societal functioning of those very same landscapes.Consequently, it is also unjust to assume unquestioningly that the early medieval wilderness still had its natural vegetation, thereby effectively denying the effects that horse keeping must already have had on the vegetation, like forest clearing and planting and maintaining thorny hedges.
All of this may sound as if deploying semi-wild horses from a conservation perspective would be something reprehensible and a bad idea.That, however, would not be a valid conclusion.One can think of many other good reasons to deploy year-round grazing as a management technique and to conserve old domestic breeds of horses and cattle.If such management techniques contribute to the conservation of biodiversity, as seems to be the case in many rewilding projects, they should be welcomed.Herds of semi-wild horses and cattle also seem to appeal to the public, and thus add to the support of nature conservation in general.It also contributes to the conservation of our cultural heritage, both by reinstating former agricultural practices -which herding free-roaming horses factually is -and by conserving rare breeds in the landscape.Provided that the reinstatement of such practices can be combined with biodiversity conservation, they should be cherished.Due to the seemingly ever decreasing space for non-economic purposes, we need to look for opportunities to combine so-called 'weak' functions, i.e. nature and the cultural-historical landscape.Deploying semiwild large grazers in nature management projects is in fact taking such an opportunity.

Translation English
If someone steals a cow or a broodmare and drives away a foal or a calf, he shall compensate [the latter to the person who is robbed] with [1] shilling, and the mother[animals] according to their value.

Comment
Especially the Instituta Cnuti-version makes clear that when spoken of equae in medieval texts, studmares are meant.

Translation English
There are also many other animals that, although living within the boundaries of the forest and subject to moderate burdens and care, can still in no way be considered belonging to the foresta, such as horses, bulls, cows and the like.

Comment
Pseudo-Cnut's Constitutiones de Foresta has been demonstrated to be a forgery (it was produced during the reign of King Henry ii of England, rather than during the reign of King Cnut), it is considered authentic to the eleventh-and twelfthcentury forest laws.
Horses (together with other domesticated livestock) apparently roamed the wilderness that was under foresta-legislation.They are, however, explicitly mentioned not to be part of the foresta.
In a footnote to this section, Liebermann further reports that in the twelfth century, Rievaulx Abbey, apart from cattle, had '60 matres equas cum nutrimento suo per forestam'.Cart.Rieval.157.67: 60 mother mares with their feeders/right to obtain fodder (?) in the forest.

Translation English
If a man has a herd of mares, he shall give a tithe of the foals; those who have only one or two, a denarius each for each foal.

Comment
This clause from the famous Leges Edwardi Confessoris makes clear what the landlord's interest was in laying down the number of untamed forest horses in the wilderness as is the case in Domesday Book: tax had to be paid on the proceeds of herds of ('wild') horses in the form of foals.As also argued in the main text of the article, this meant that the herds did have owners and it had to be known who those owners were, otherwise taxes could not be levied.bibliography Liebermann, F., 1903. Die Gesetze der Angelsachsen. Erster Band. Text und Übersetzung (Halle, Germany).Molyneaux, G., 2012.The Ordinance concerning the Dunsaete and the Anglo-Welsh frontier in the late tenth and eleventh centuries ',40,

Fig. 1 .
Fig. 1.Distribution of equae indomitae and equae silvestrae/silvaticae/ silvestris as listed in Domesday Book.The shires of Domesday England are indicated in light brown.Medieval and early modern mentions of 'wild' horses from other sources are shown as black dots.Domesday Book mentions equae silvestrae as well as equae silvaticae and equa silvestres.Because the distinction is not always clear (due to abbreviations used in Domesday Book), these three names have been combined here.Map created with QGIS, QGIS.org (2022).QGIS Geographic Information System.QGIS Association.http://www.qgis.org.

Fig. 2 .
Fig. 2. Distribution of different types of 'wild' horses in Europe during the Middle Ages and Early Modern Period.Mares are shown as circles; 'neutral' horses are shown as triangles.It cannot be excluded that the latter category also consisted entirely or largely of mares.Map created with QGIS, QGIS.org (2022).QGIS Geographic Information System.QGIS Association.http://www.qgis.org.

Fig. 3 .
Fig. 3. Distribution of haiae and parci as listed in Domesday Book.The shires of Domesday England are indicated in light brown.For reference Domesday Book 'wild' horses as well as other medieval and early modern mentions of 'wild' horses are shown as grey dots.Map created with QGIS, QGIS.org (2022).QGIS Geographic Information System.QGIS Association.http://www.qgis.org.
acknowledgementsI would like to thank my friends and (former) colleagues Prof. Mark Huijbregts and Dr Thomas van Goethem for their constant encouragements to publish this research.My gratitude also goes to my former students Marc Stadman and Gemma Tuckwood for their help in preparing the manuscript.
are convinced they did survive.Evidence for this, they claim, can be derived from Domesday Book.Descriptio totius Angliae) is the record of a survey commissioned by William the Conqueror of much of England and parts of Wales completed in 1086.One of the main purposes of the survey was to determine who held what in loan and what taxes were consequently liable.Among the items listed are surface areas of arable land, meadows, pastures, and woodland, as Abstract of Claims Preferred at a Justice Seat Held for the New Forest, Hants, in the twenty-second year of the reign of King Charles II.A.D. 1670 (London).Ahrland, Å., 2011.'Power and paradise.Swedish deer parks in a long-term perspective', Bebyggelsehistorisk Tidskrift.Nordic J Settlement Hist & Built Heritage, 61, pp.68-89.Bendrey, R., 2012.'From wild horses to domestic horses: a European perspective', World Archaeol, 44, 135-57.Bestohlene] mit [1] Schilling und die Mutter[thiere] je nach ihrem Werthe.