Explaining Anglo-Saxon military efficiency: the landscape of mobilization

Abstract The importance of warfare in Anglo-Saxon England is widely accepted, but the processes by which armies were put in the field are only partially understood, with most discussion focusing on the economic logistics rather than the spatial practicalities of mobilization. Yet such a system underpinned recorded military actions and must have evolved in response to changing military organization in the late Anglo-Saxon period. Through an assessment of documentary references to sites of muster, and by using a multidisciplinary landscape-focused approach, this article examines possible traces of that system – especially those preserved in place-names – and relates them to later Anglo-Saxon administrative geography.

known that they could be used by troops travelling from more than 80 km away; they were also presumably distinctively marked in the landscape -perhaps monumentalized as the name Egbert's Stone suggests -to allow their identification locally. It is also implicit from this account that mobilization relied on efficient systems of intelligence and communication. Information about when and where to muster was managed to such a degree that Alfred only needed to stay one day at Egbert's Stone; any provision made for latecomers had a tight timeframe, since Alfred went into battle just two days after arriving there. A further implication is that mustering sites were located on significant long-distance routes, accessible, in this instance, to troops from three shires, and appropriate for the massed forward movement of the armies to engage the enemy. According to this text, such musters were apparently connected to the cadre of fighting units: the shires. Finally, it is clear that the encounter with the Vikings was carefully coordinated. Alfred's staged movement northwards, and the Great Army's advance from Chippenham culminating in their encounter at Edington could only have been achieved either by bilateral agreement or at the very least very careful observation (leading to ambush) of one army already intent on aggression by another that was keen to fight. Presumably because of the difficulties in assembling and maintaining the coherence of armies, pitched battles on open ground are extremely rare in the sources. The detail of the account in the Chronicle makes clear that this was a carefully orchestrated encounter.
The above observations suggest that, whilst the mobilization of Egbert's Stone may have been exceptional, the system on which it relied was not. Moreover, although references to the assembling of troops in Anglo-Saxon England may on the whole be rare, several others can be identified. Bede mentions a mustering at Wilfaraesdun in 644 in preparation for a battle that never took place, 3 while Aylesford (1016), Beverston (1051) and Langtree (1051) are all named as such in the Chronicle. 4 The meeting at Swinbeorg, mentioned in Alfred's will, may also be relevant if, as suggested by Marren, it was the site of a muster prior to the battle of Ashdown in 871. 5 However rare explicit references to mustering sites may be, working systems of mobilization are implicit in our understanding of Anglo-Saxon military organization, and underpin many of the events recorded in the Anglo-Saxon Chronicle and other narrative sources. The narrative of AEthelmund's attack on Wessex in 802, for example, which was met before the Hwiccan force could advance beyond the Thames, highlights the speed with which an effective local militia (in this instance the Wilsaete) could be assembled, even during an interregnum. 6 The conduct of all warfare, as von Clausewitz makes clear, is predicated by the assembly of forces in time and space. 7 The nature of such assembly would, however, largely depend upon the kinds of armies being raised, the form of the conflict, and the systems by which troops were levied. All three of these areas underwent significant changes during the period, with the later ninth and early tenth centuries in particular marking a point of considerable innovation in military strategy in conjunction with an escalation of warfare. These changes will have impacted on the requirements and logistics of manpower, and, therefore, on systems of mobilization.
A first key distinction can be drawn between forces assembled for offensive and defensive campaigns. 8 It has long been argued that during the early and between the West Saxons and Viking invaders; so the probability that the Swinbeorg assembly acted as a muster prior to battle is high. 6 Table 5.2, helpfully sets out probable average speeds for different elements of an army, some of which were mounted, but makes the point that 'if 20 miles (31.2 km) was a long day's march, 10 miles (16.1 km) may be reckoned to be the maximum distance after which a force could still be expected to be in condition to fight after travelling'. Since the battle of middle Anglo-Saxon period the former consisted essentially of elite warriorswhat Hollister conveniently labelled the 'select fyrd' of kings, the king's landed dependants, and other land owning lords, along with their household followerswhich in the Anglo-Saxon Chronicle are always described by their geo-political affiliation (that is, West Saxon, Mercian, and so on). 9 These can be contrasted with defensive forces, which narrative accounts suggest were of regional identity and ad hoc mobilization. 10 Examples of this type include West Saxon responses to armed invasions in the ninth century which seem to have been based on shire units (the Wilsaete in 802, Sumorsaete and Dornsaete in 848, the men of Hampshire and Berkshire in 860); often led by the local ealdorman; and assembled for a specific task (a single battle or rapid succession of battles). The precise composition of these armies is much debated, but the assumption -encapsulated most clearly in Hollister's concept of the 'great fyrd' -is that these consisted of much larger forces, drawing (in addition to the select fyrd) on militias of freemen. 11 The shire levy, as Powicke points out (and as Alfred's Laws make clear), was not only the first point of reference in policing local disturbances, but a basic unit from which larger defensive forces could be constructed. 12 If the shire was one of the systems through which armies were raised, it is clear that developments in the eighth and early ninth centuries more closely linked together land owning and military service. It is at this time -certainly in Mercia, and probably in Kent -that charters with  10 Ninth-century descriptions of the armies of the sixth to eighth centuries may conform to later Anglo-Saxon preconceptions, but those of the late eighth and ninth centuries would have been assembled within the living memory of, or at only a few removes from, the writers of late-ninth-century texts. 11 On this, and the subsequent point, it is significant to note that unlike recent historical writing, which has downplayed the role in civil defence assigned to the common citizenry (e. clauses reserving military obligations provide the first evidence of more systematic military structures being in place. 13 These formalized, on the one hand, the nature of levies and the obligations imposed on the landowning sections of society, and, on the other, the physical infrastructure of military authority. Against this backdrop it is easy to envisage the operation of two parallel systems of mobilization, defined according to their defensive or offensive purpose. Aggressive campaigns, with forces often led by kings and consisting primarily of warrior elites, which by their nature permitted a greater level of planning and logistical support, are likely to have used networks of royal and aristocratic landholding in order to mobilize. 14 By contrast, emergency defensive operations, which drew upon a wider population beyond the warrior elite, would have been more closely tied to existing civil structures with predetermined mustering places. 15 On the other hand, there is no reason why the personnel of offensive or defensive forces should have differed significantly in terms of socio-economic status. The actual fighting force in each case may well have consisted of the same thegnly cohort; but a campaigning army is unlikely to have been operational without the support of a significant non-combatant contingent. 16 Crucial differences can be assumed in the size of the forces mobilized (including both armed and unarmed participants) and, more importantly, in the processes of mobilization involved. In the first case, numerical variance would be a natural corollary of the raising of forces from a kingdom-wide pool on the one hand, and a single shire on the other (even if somewhat counterbalanced by a more selective wide-scale mobilization than was expedient locally under emergency conditions). Differences in the processes of mobilization would follow not only from this, but also from the different requirements of the carefully planned constitution of an expeditionary 13  force for a long, external campaign, and the rapid mustering of troops to confront an immediate, possibly unpredicted and temporary territorial threat. Places of assembly were important not only from the perspective of efficient mobilization, but also from that of strategy more generally. Certainly, efficiency in response to threats depended on intimate knowledge of the communications infrastructure. It may also have been due in part to a system of early warning -there is clear evidence for the existence of lookouts across southern Britain. 17 However, neither of these would have been of much use without a well-choreographed process of mobilization, allowing troops to be raised and manoeuvred into the required position in sufficient numbers and at adequate speed, without simultaneously being kept in the field longer than their motivation or provisions permitted. 18 This concern would have been acute even for armies made up only of eligible members of the weapon-carrying elite. In some analyses, shire militias depended on the mobilization of as many able-bodied freemen as possible in the areas nearest to the perceived threat, representing, therefore, a serious temporary drain on agricultural manpower. This too would have limited the time a force could be kept in the field and increased the importance of rapid mobilization and clearly defined responses.
Alfred's reforms, recounted in the Chronicle entry for 891, can be seen as an attempt to enhance the role and effectiveness of central authority in these mobilizations. The division of the fyrd into two 'so that always half of its men were at home, half on service, apart from the men who guarded the boroughs', 19 though ultimately designed to provide for greater continuity in military actions, also required more persistent supervision and administration. The rotation of troops stationed in strongholds with those on a tour of duty could not have been achieved without advanced administrative efficiency, 17  and nor could the logistical concerns of supplying simultaneously stationary garrisons and mobile forces. Further administrative complications can also be surmised regarding the raising of armies, and the methods of calculating duty and its cash equivalent. 20 Since the final local muster in any such mobilization was presumably the burh, 21 these reforms may over time have replaced the pre-eminent shire or multiple-shire mustering points, such as Egbert's Stone, in practical importance. Mustering within a burh removed some of the dangers of fragmentation, providing supplies, security and spatial coherence to the mustered force.
Whether discussing the composition of armies and their military goals, the territorial basis of military service, or the logistical concerns of mustering, campaigning, and battle, the crucial geographical framework for the bringing together of forces -the landscape of mobilization -is often overlooked in studies of Anglo-Saxon militarism. While the processes may be obvious -the gathering of troops at pre-arranged mustering sites -the sites themselves have much more to reveal about the composition and structure of early medieval armies. This paper makes use of landscape evidence to advance our knowledge of Anglo-Saxon mobilization beyond the limits of narrative, administrative and fiscal sources. It considers the link between known assembly places and systems of muster, and suggests a number of other sites at which military gatherings may have taken place. Furthermore, it argues that changes in the organization of armies over the course of the later Anglo-Saxon period would have had an impact on the mechanisms that permitted military obligations to be carried out, and this impact is likely to be detectable in the organization of the landscape of mobilization. 20 In one of the most detailed discussions of this relationship between land tenure and military service, Hollister Anglo-Saxon Military Institutions, pp. 38-58, has drawn particular attention to the emergence by the eleventh century of the 'five-hide unit': a rule of recruitment visible in most of the shires of southern England, by which every five hides of land produced one warrior for military expeditions. However, the five-hide unit was a fiscal as well as a military assessment; in the absence of a suitable warrior-representative, service could be valued at the equivalent rate of 2s on the hide (ibid. p. 48 Cuckhamsley Knob (Cwicelmeshlaewe) fulfils all of the criteria described below for a here-hlāw mustering point. The failure of the vikings to provoke a response from the English by occupying this location might simply signify that this shire-level system had already been superseded by that of burghal mobilization, at least in Wessex, depriving Cwicelmeshlaewe of its military functions (though not necessarily its legal and administrative ones); the vikings were perhaps standing on the wrong spot! We have used the term 'burh' in normal type and treated it as a part of modern vocabulary, rather than an Old English term, in order to reflect its use as a normal part of modern archaeological and historical discourse. We recognize that OE fyrd and here sometimes fall into the same category; here they are given in italics to avoid confusion with modern English 'here'.
tHE HunDrEDAL systEM AnD MobiLizAtion In late Anglo-Saxon England, one highly organized system that regularly brought large numbers of people together is very clearly attested -the network of hund reds or wapentakes into which each shire was divided for administrative purposes. In all likelihood, places of legal or governmental assembly were an integral part of the Anglo-Saxon system of mobilization, and indeed several of the musters recorded in the Chronicle have been identified as the centres of Domesday administrative districts. 22 That this system of judicial governance intersected to some degree with that of mobilization would hardly be surprising: as the socially ingrained sites of regular mass gatherings, sites of judicial or governmental assembly -the meeting-places of hundreds, wapentakes, lathes and so on -would have formed a reliable basis for military musters; indeed, it might be considered that that is precisely what they originally were, since many of those attending such assemblies belonged to the weapon-bearing sections of society. Moreover, as military recruitment was assessed locally on units of landholding, it is probable that mustering -whether for campaigning or to review the completeness and order of troops under arms -required mechanisms which were closely related to those of taxation, administration, and judgement in the localities. The principal institution fulfilling these roles, at least by the mid-tenth century, appears to have been the hundred. How clearly these apparently separate functions were differentiated in early times, or where the original emphasis of such meetings lay, are difficult questions to answer, but some underlying military function can be discerned. 23 Ealdorman AEthelweard, writing in the later tenth century, seems to endow the term 'hundred' (centurias) with military characteristics. 24 The hundredal system as it appears in Domesday is superficially irregular and had apparently already undergone several phases of modification by the later eleventh century; but the archetypal system that underlies it was clearly both highly regular and tightly regulated. Legislation decreed that hundred meetings should be held every four weeks and that all freemen should attend. 25 Even the name of the administrative units, which notionally consisted of 100 hides each (whether or not they did in practice), suggests regularity and, perhaps, top-down imposition, at least in parts of England (and perhaps following a continental exemplar). 26 A connection might be made between this and other very regular elements of the late Anglo-Saxon military administrative system -the five-hide unit of military and fiscal assessment, the three-hundred hide ship-soke, shires (at least in parts of England) apparently arranged in a geographically regular way around a burh, and so on. 27 The hundredal system should also be seen within the context of recruitment to the army, for by the late ninth century this too was very regular. Under Alfred, to reiterate, West Saxon human resources were divided into three parts -those who garrisoned the burhs, those who served in the fyrd and those who tended the land -and importantly, the fyrd was always in service. This could be achieved in more than one way, and perhaps the staffing of these shifts involved massive periodic changes in personnel -this is certainly implied by the Chronicle's comment that the fyrd besieging Haesten at Thorney had completed its tour of duty (hie haefdon þa heora stemn gesetenne) before Alfred could arrive with a replacement force. 28 Nevertheless, regularised military service would be harder to manage if the manner and frequency of musters and the numbers of troops mobilized varied too much from district to district. 29 It is tempting therefore to posit the establishment of a relatively regularly structured system of multi-functional moots/military musters more or less contemporaneously with and supplementary to a similarly regular system of fortification and military organization. As Molyneaux notes, it may not be a coincidence that we first hear of the hundreds during the tenth century. 30 Even if we accept that the hundredal system, and the meeting-places of those administrative districts, formed the basis for an efficient and regular system of mobilization in late Anglo-Saxon England, a number of important gaps remain in our understanding of the landscape of mobilization. It is very hard to be sure if any of the district assembly-sites fossilized in the Domesday record go back beyond the late Anglo-Saxon period, and since most of the musters mentioned in the narrative sources date from the age of Alfred or later, pre-Alfredian sites of muster are more or less invisible in documentary terms. Yet kings had been emphasizing the central importance of public assembly at least since the seventh century, 31 and reserving military obligations since the mid-eighth; 32 so the centrality of public assembly at a much earlier period is strongly hinted at, and we know that warfare and the exercise of military authority were integral parts of society in the early and middle Anglo-Saxon periods. It seems very likely, therefore, that elements of the hundredal system were in place from an early period -existing perhaps within a very different administrative framework -and that at least some of their functions were 28 ASC MS A s.a. 893; Anglo-Saxon Chronicle MS A, ed. Bately, p. 56. 29 A more efficient system is easy to imagine, and would involve almost continual and regular replacement of troops, so that the coherence of the army was maintained even as the personnel changed; in other words, a system of mobilization that almost precisely reflects the regularity that the hundredal system provided. 30 G. Molyneaux, 'Why were Some Tenth-Century English Kings Presented as Rulers of Britain?', TRHS 21 (2011), 59-91. It may also be significant that the system of districts actually called hundreds extended only as far as the midlands in any systematic way, and beyond that gave way to wapentakes; it might only have been in southern England that late Anglo-Saxon kings were able to levy forces in this regular and systematic fashion. 31 AEthelberht, ch. 1, and Hlothhere and Eadric, ch. 8 military. Establishing which sites formed the basis of mobilization prior to the Alfredian and later reforms is of considerable potential importance for our understanding of the equivalent arrangements in, say, Mercia and Wessex. It is worth noting, moreover, that later Anglo-Saxon tactical units at the level of the shire, burh, and ship-soke may also have required identifiable places of assembly. The origins of these groupings are somewhat unclear, 33 but it is likely that developments in military organization were accompanied by changes in the organization and the landscape setting of musters. A careful, landscape-based approach may be able not only to demonstrate the existence of pre-hundredal elements of a system of mobilization, but also to throw light on the potentially hierarchical nature of later Anglo-Saxon arrangements.
tHE LAnDscApE AnD LAnguAgE oF MobiLizAtion A landscape approach may be able to supplement our knowledge and understanding of this important aspect of Anglo-Saxon warfare. Here we meet with an immediate difficulty, for sites of occasional temporary use are unlikely to leave a marked and easily identifiable imprint in the archaeological record. 34 Topographically, it is easier to define likely contexts for military gatherings than actually to identify sites of muster themselves. For example, major gatherings of this type -shire or multiple-shire musters of the Egbert's Stone type -are likely to have been easy to access, perhaps from more than one district or shire, and well-marked on the ground, so proximity to important route-ways, centrality to areas of settlement, and the presence of a defining natural or artificial feature should be anticipated. Whether or not Anglo-Saxon military mustering has left anything other than the most ephemeral of physical imprints, the recurrent 33 Indeed, these districts are likely to have originated at different times: the main shires of Wessex make their appearance in the sources during the eighth and ninth centuries, burhs during the ninth century, and hundreds and perhaps the ship-sokes in the second half of the tenth century: S. gathering of large groups of armed men, and the specialized function of the sites at which they came together are likely on occasion to have left a long-standing mark in local toponymy, above and beyond the names of mustering sites discussed above, and a consideration of certain types of place-name may therefore provide a means of identifying some of the sites at which musters were held. The names of the small number of recorded mustering sites noted above are worth consideration. Ecgbrihtesstan, Beverston, Aylesford, Langtree and Swinbeorg have the generics stān 'stone', ford 'ford', trēow 'tree' and beorg 'mound', and of the Domesday hundreds named after their meeting-places, almost two-thirds are named after stones, mounds, trees or river crossings. 35 The names of most of these mustering sites conform in one way or another with the types of compound regularly associated with the meeting-places of late Anglo-Saxon administrative districts. Their first elements, on the other hand, less clearly belong to a single type. Personal names dominate (if that can be said of such a small corpus), forming the first elements of Aylesford, Ecgbrihtesstan, Wilfaraesdun and perhaps Beverston, and although some of those names might have resonated with historical or mythical significance -Egbert was an important ancestor of Alfred and AEgel was a figure from Germanic legend -others did not, or at least not obviously, and the significance of such first elements is difficult to assess. 36 This corpus might be supplemented by examining place-name terminology descriptive of military activity, and in particular the Old English (OE) element here. The main meaning of here is 'an army', and although a more general application to 'a multitude of people' seems to have developed, the sense 'army' is shared by its Germanic cognates and must be treated as primary. 37 Although OE here is well represented in place-names, it is striking that very few hundreds or their identified meeting-places have names of this kind. 38 The possibility that the hundredal system intersects with a system of mobilization is discussed above, yet the toponymic evidence for military activity at such sites is only slight. Among a class of sites whose principal function was the bringing together of significant bands, it may not appear surprising that names indicating the assembling of an army did not seem very distinctive; but place-name references to judicial activity and debate -other obvious functions of these sites -are relatively common by comparison. 39 OE here is particularly well represented in three recurrent place-name compounds (Fig. 5): OE here-hlāw 'army mound', here-feld 'army open land', and here-ford 'army ford'. 40 These compounds are found in a considerable number of place-names, and this suggests that they constituted functional labels -the language of administrators, attached to place-names and later fossilized as place-names proper (analogous with OE āc-tūn or burh-tūn) 41 -rather than ad hoc toponymic formations in the language of local inhabitants.  43 Of about 800 hundred-and wapentake-names, 33 bear the generic hlāw, 19 the generic feld and 40 ford. 44 In other words, although this group of lexical compounds may not be found very frequently in the names of hundreds themselves, they clearly share in the language of assembly. The complete lack of similar types of compound in OE folc or maegð (both of which mean 'folk, people' and both of which occur in place-names and compound nouns) is also telling. It suggests, for these compounds, the active choice of a word denoting a fighting force rather than simply a large body of people. 45 There has been little discussion of here-hlāw as a recurrent compound. Ekwall took the first element of Harlow (Essex) to mean 'the whole people', perhaps principally because this was also the name of a Domesday hundred; 46 but the lack of any close correlation between Harlow names as a group and known hundred meeting-places, and the general rarity of hundred-names referring to multitudes of people, seem to count against this explanation. 47 Discussion 42 The problem is twofold. Does here in such place-names refer, specifically or figuratively, to an actual army or to a large group of people? Is the sense literal -'frequented by an army or a large group of people'; metaphorical -'suitable for an army or large group of people'; administrative -'maintained by the nation or army'; or general -'public, used by or associated with everyone'. It might even be suggested that names of this kind record a particularly famous event involving an army or large group of people -a battle perhaps (cf.  48 and in the final edition of his dictionary, Ekwall still felt that 'the meaning of such a name is not apparent', although he raised the idea of assembly of some kind. 49 Mills suggested that the 'army' signified by the element here was perhaps a Viking army, 50 probably because Alfredian narrative sources distinguish between the Anglo-Saxon fyrd and the Viking here; but this is far from convincing, not least on grounds of distribution, and Watts says that the 'reason for the name is unknown'. 51 OE here-ford is another traditionally troublesome compound in toponymic discussion, but the corpus is fairly well established. Attempts to explain the compound have proved difficult. Stenton and Mawer provided a translation without further comment: 'army ford'. 52 Ekwall took here-ford to denote a ford 'where a marching column could pass in closed order', 53 a sense closely echoed by Mills, 54 but Torvell dismissed any military relevance in most of the place-names, pointing out that at least one of the fords (Torvell's identification of the likely ford at Harford, Devon) could never have been appropriate for the crossing of an army. 55 He felt also that such names could not be the commemoration of single famous military campaigns, which might be expected to produce more specific names such as 'Hengist's ford' or similar. Instead, in Torvell's analysis, a here-ford was more likely to have been a river crossing of regional importance, specifically maintained by the local inhabitants, so that it was passable all year round, either because it was on a route of major economic importance, or because other routes would become treacherous during certain seasons, forcing travellers to pass by the here-ford. However, this explanation also requires a certain amount of special pleading regarding the economic significance of the fords and route-ways so named and of their settlements, given the small number of surviving examples.
Alternatively, there may be some connection with the construction herepaeð-ford, found at least nine times in place-names and charter boundaries: five times in Devon, twice in Somerset, and once each in Oxfordshire and Surrey; but here-ford is unlikely to be a reduced form of the triple compound. While it is clear that here-paeð-ford names tend to undergo reduction over time, a reflex of the middle syllable is generally still evident until fairly late -up to the present time in modern forms such as Hartford and Harpford, and into the fourteenth century at least in the case of Harford in Devon. 56 The survival of Domesday and pre-Domesday forms for the here-ford names requires such a reduction to have occurred at a very early stage. On the other hand, here-ford might be elliptical for here-paeð-ford, where here has become spoken shorthand for here-paeð. 57 The distributions of here-ford and here-paeð-ford overlap, in as much as the latter is found concentrated in southern England, between Hartford in Devon and the remaining here-ford names which are north of the Thames. However, the distributions are essentially different, raising the possibility of a regional variation. Nevertheless, this solution is problematic and Carroll and Parsons consider it to be uncertain. 58 Since the same compound occurs in continental toponymy, 59 it might be necessary to posit parallel early developments of here-paeð-ford and an elliptical here-ford across the West Germanic language region, and this seems an unsatisfactorily convoluted means of explaining the compound.
The possibility that here-ford might refer to a site at which large gatherings took place does away with any need for the fords in question to be communication routes of national importance, or for the roads on which they lie to be of here-paeð type or to be major long-distance route-ways of any kind. A here-ford, even as a place of military muster, would not need to be situated on a defensive alignment. We might expect places of muster to be in or close to strategically important zones, but not necessarily in positions of strength. A ford marking a site of assembly need not be a major ford (and indeed frequently is not), since it only needs to mark the point on the road where the meeting is going to take place -as long as it is distinctive enough to serve that purpose, and providing its surroundings meet any practical requirements, it can be almost insignificant as a crossing point or infrastructural hub. Finally, a mustering-or meetingplace might well occur on a major route-way, but could just as easily be on a smaller road linking other elements of the transport network. Herford. 60 The ford at which the hundred of Uttlesford (Essex) met, now marked by Uttlesford Bridge, was probably a minor river-crossing by a major road; that of Holeford hundred (Gloucestershire) was distinctive, but probably not significant; and Barford (Bedfordshire) The interpretation of herenames proposed here removes a major stumbling block in the understanding of the names as a class. If they are functional appellative compounds, there is little need for the landscape features to which they are attached to share precisely the same physical characteristics; it is their role in the organization of society that unites them, and therefore their wider landscape is perhaps more significant than their immediate topographical form. Given that the primary sense of OE here is 'army', the most natural interpretation of these three compounds is as sites where armed groups gathered periodically. That such sites might also have held other administrative functions is quite possible, but would not diminish the force of this argument. tHE topogrApHy A select corpus of place-names with forms likely to represent OE here-hlāw, here-feld or here-ford, and related compounds, is set out in the Appendix. These names belong to an administrative language of assembly, and make specific reference to armed groups. Their topographical locations suggest that they also form part of a landscape of assembly. Many of the sites display one or more of a range of features shared with known assembly places: proximity to important route-ways; association with boundaries of some significance; affiliation with monumental sites or places of early ceremonial importance (which may in some cases predate administrative arrangements of the later Anglo-Saxon period). Moreover, several instances of the compounds discussed above have a spatial relationship to assembly sites identified from other evidence.
The only here-hlāw place-name with pre-Conquest forms is Harlow in Essex (1), 61 which gave its name to Harlow half hundred, so its credentials as a meeting-place are reasonably secure. 62 The meeting-place of the hundred is thought to have been an earthwork mound c. 25m in diameter and 1.5m high located on the edge of Mulberry Green. 63 This was Mudborow, Mudbrowne Grene met at the crossing of a major river, but not on an obviously important early long-distance route-way. 61 Numbers in brackets refer to the list of herenames in the Appendix. 62 Christy, Gelling, and Watts take the name to refer to the small hill west of the railway station, on which a Romano-Celtic temple stood, but Ekwall is surely right in taking the name to refer to the hundred meeting-place, which is focused on a still-visible mound. The element hlāw normally denotes artificial mounds in southern counties, and Gelling explains the posited reference to the natural hill of Harlow temple on the basis that part of this hill had been artificially scarped; but the presence of another, wholly man-made mound provides a more con- in the seventeenth century and Moteberghe in 1382, later reflexes of OE (ge) mōt-beorg 'assembly mound'. 64 This name may only have arisen after the placename Harlow had become semantically opaque, and it certainly emphasizes the tradition of gatherings on the site. Perhaps indicative of this function, the mound has a slightly flattened summit of c. 8 m diameter; a feature common to purpose-built assembly mounds in England and Scandinavia. 65 There is no known record of assembly at Harlow in Staffordshire (9), but the place lies in Mayfield parish. 66 This has been derived from the plant-name madder with OE feld, 67 but Duignan and Horovitz suggest an alternative first element OE maeðel 'meeting, council', 68 which certainly seems to explain some of the forms more satisfactorily. 69 Harville (Kent) is in Wye parish, located approximately 1 km west of the centre of Wye, the head not only of a Kentish hundred, but of the Domesday lathe of that name. 70 Given this association it is likely that meetings of hundred and lathe took place close to the settlement 64   and therefore very close to Harville. Few other sites discussed here produce such clear evidence of communal gatherings, but a tradition of assembly might be assumed at Hereford (14), which became the centre of a shire, while Hartford (17) is just 2 km north-east of Huntingdon, another shire caput and a late Anglo-Saxon burh. The meeting-place of Elthorne Hundred, the district in which Harefield (10) is located, has not been satisfactorily identified; 71 it is not impossible that the thorn-bush alluded to in the name was situated somewhere in the here-feld. The spatial association of about a third of the corpus with Anglo-Saxon meeting-sites attested on historical grounds or through use of independent toponymic evidence is sufficient to suggest that this class of place-name might be connected with assembly in some way. 72 In common with attested assembly sites, many of the here place-names also show close proximity to elements of the middle-and long-distance transport system of the Anglo-Saxon period. The three Essex Harlows are all on or close to important Roman or prehistoric routes: the Roman road leading to Braughing (Mgy 329) (1), the Icknield Way (3), and the Roman road from Braintree to Bury St Edmunds (Mgy 33a) (7). In fact, Harlow (1)  The association between hereplaces and major shire and regional boundaries is emphasized by several further examples. Harlow in Mayfield (St) (9) lies just 1.2 km west of the Domesday border of Staffordshire and Derbyshire. Little Hereford (He) (15) is 300m north of the Herefordshire/Worcestershire boundary, and is only 2.8 km west of the point where the boundary of those two shires met that of Shropshire. Similarly, Harvington in Worcestershire (13) lies 1 km south of the boundary with Warwickshire, and 5 km west of that with Gloucestershire. Finally, Harefield (Mx) (10) lies 1.7 km away from the threeway intersection of Middlesex, Hertfordshire, and Buckinghamshire.
If these place-names do reflect the existence of mustering points, then the material discussed so far might not be comprehensive. OE here-wīc '(dependent) settlement associated with an army' occurs in at least three place-names in Kent, Essex and Lincolnshire, and there have been suggestions that these might have been military encampments. 78 The compound is analogous with the attested fyrd-wīc, which is very clearly used of military camps. 79 The coastal location of Harwich in Essex and Harwich Street in Whitstable, Kent, might have been suitable for a naval muster, or a mustering prior to embarkation. Whitstable, it should be noted, was the centre of another Kentish hundred. Together with OE here-ford we should perhaps also examine OE þēod-ford, a compound identified in four East Anglian and Lincolnshire place-names. 80 Although the military connotations of OE þēod 'a nation, people' or 'a district' are less obvious than for OE here, 81 the evocation of the involvement of a large group of people (perhaps a group of people representative of a nation), or of association with an established district, is certainly in keeping with the function of a mustering site. Indeed, a military association has been noted. 82 Although þēod is taken to mean 'public' in some instances, 83 and might well have that sense when compounded with generics denoting route-ways or features of route-ways, such as fords, it is still necessary to explain the specific significance of such compounds, and to define the ways in which they were considered to be especially 'public'. Since the compound occurs only in one region, and does not apparently overlap with the semantically comparable here-ford, it might be proposed that the two compounds denote similar sites. It is worth noting that Thetford in Norfolk was a hundred meeting-place situated on the boundary between Norfolk and Suffolk, at the crossing of the Little Ouse by the Icknield Way and at the convergence of two important Roman roads: Mgy 333 and Mgy 332. 84 Perhaps surprisingly, OE fyrd 'army (service), military expedition', the term frequently used in narrative sources to describe the West Saxon and Anglo-Saxon armies of the ninth and tenth centuries, is relatively rare in place-names, found most often in compound with generics referring to roads, and never apparently in compound with hlāw, ford or feld. 85 Three instances with other generics are worth noting: Verdley in Sussex (Verdelay 1318), where the second element is OE lēah 'woodland' and an (ealden) fyrdgeat mentioned in the bounds of two separate Anglo-Saxon charters, for Headborne Worthy and Micheldever in Hampshire, 86 but probably referring to a single feature. 87 As Mawer et al. note Verdley may have 'taken its name from some long forgotten hosting'. 88 In the second instance we have a compound of OE fyrd and geat 'gap, gate', with OE ald added in one instance, hence 'the (old) army gate'. As many as seven hundreds seem to have met at places with names in geat, suggesting that gates were somehow appropriate markers for meeting-places. 89 If Grundy's identification is correct, the site is close to the Winchester to Andover Roman road, and it is worth noting that Micheldever was a villa regalis. The third instance is the fyrdhammas in a charter for Stanmore, Berkshire. 90 The second element is clearly OE hamm (pl. hammas), but there is uncertainty surrounding the first. 91 Gelling took it to be OE fyrde a variant of ford, hence 'ford meadows', but acknowledged that this variant was not otherwise found as a first element. Others have taken the first element to be OE fyrd, 92 in which case the compound might mean 'army enclosures', and Forsberg posited a mustering-place for the Berkshire fyrd, though preferring to explain fyrd in this and the previous instance as ellipsis for fyrd-straet or fyrd-wīc, hence 'enclosures by the fyrd-straet'. The only known compounds of fyrd with straet or other words meaning 'road' are located in the midlands, 93 while here-paeð is well-evidenced in Berkshire, so it seems unlikely that an elliptical reference to a fyrd-straet should occur in this area.
Whether or not þēod and fyrd place-names are included, the material discussed above clearly raises a number of questions; two are particularly important. Firstly, does the geographical positioning of the sites have significance, both in terms of their individual landscape settings and of their overall distributions? Do these allow us any level of confidence concerning the original function of the sites? The etymological background of the place-names is certainly in favour of reference to military activities and specifically to the mustering of armies, but such an interpretation requires external support. Secondly, if these are mustering sites, when were they in use as such? This question may be fundamental to understanding the origins and structure of Anglo-Saxon administrative districts, the hierarchical relationship between different types of meeting-or mustering-place, and the workings of mobilization, and should perhaps be discussed first.

DAting oF HErE nAMEs
It is notoriously difficult to deduce dates for the first coining of individual place-names. Usually it is possible only to state a date by which the name was in use, which is the same date as its first record. On that basis we cannot push the date of any here name discussed here far back into the Anglo-Saxon period. Harvington is recorded in charters purportedly dating to 709, 799 (for 802?) and 963-4, 94 but none of these survives in a manuscript copied earlier than  , 1972), no. 10), S 154 (from Worcester, probably authentic, but with a question over its date; the earliest surviving copy is in an early eleventh-century manuscript, but an incomplete seventeenthcentury copy of the lost original also survives, e. g. Finberg, West Midlands, no. 233), S 731 (a the eleventh century and all are of disputed authenticity. On the other hand, the names in question could have been formed decades or centuries before their first surviving attestation. The absence from early texts (especially from narrative ones) of a class of compounds that makes reference to large groups of people, whatever their reason for gathering, is perhaps unexpected given the many references to events that would have involved major gatherings -grand diplomacy, synods and battles -but not entirely surprising if compounds such as here-hlāw were functional appellatives rather than organic place-names. 95 Another possible clue to the dating of this class of place-names can be taken from their generics. While a number of them are compounded with OE ford, a similar compound in OE brycg is extremely rare. 96 The latter element could mean 'causeway' as well as 'bridge' and so need not belong to a period of advanced bridge technology, but it is worth noting that properly maintained bridges do not seem to have become a widespread feature of Anglo-Saxon England before the middle of the tenth century. 97 While ford is the generic in 40 English hundred-and wapentake-names, brycg occurs 18 times and is therefore a significant generic for districts or their meeting-places. Some of these may have originated as ford names, becoming brycg names only when a bridge replaced the ford, so there is no need to assume that they are part of a late group of meeting-places; but the relative absence of references to bridges among here names may indicate that they were long-fossilized place-names by the time bridges were commonplace (and perhaps by the time bridges replaced the crossings they named).
OE here-ford, here-feld and þēod-ford all have exact continental parallels in  100 What this tells us about the chronology of such compounds is unclear. It is perhaps suggestive of a shared language community, which would be consistent with an early datebut most of the examples are found in areas where branches of West Germanic very close to English were spoken, and where similar compounds might have evolved collectively or independently. The elements in question remained active lexical items on the continent and in England throughout the period. 101 It may also be significant that fyrd is so rarely found in this type of compound. There are no known instances of *fyrd-hlāw, *fyrd-ford or *fyrd-feld, although of course the compounds fyrd-straet and fyrd-weg do occur, and the present survey has only uncovered two or three reliable instances of fyrd in compound with elements that do not refer to route-ways. Old English fyrd seems to be unique among its Germanic cognates in its semantic evolution. The root meaning of Proto-Germanic *fartiz is likely to have been something like 'journey', and this sense is apparent in most branches of the Germanic language group. 102 Only in Old English does this root sense seem to have developed into '(military) expedition', 'military service' and ultimately '(expeditionary) military force'.
Elements cognate with fyrd are found occasionally in continental placenames, but with connotations of travel or movement. For example, Vaartbroek near Eindhoven in Noord-Brabant, the Netherlands (Vaart Broek 1837) is taken to be a reference to 'navigable marshland', 103 and Förstmann renders OHG fart, OFris ferd in place-names as 'iter, via'. 104 Of course, English place-names containing fyrd might have the same sense, but this seems unlikely if compounds such as fyrd-weg are taken into account, since it would give a meaning 'journey/ expedition road' -in other words a semantically superfluous specific. Indeed, this sense is not recorded in Old English texts (DOE sub fyrd) and may therefore have become obsolete at an early stage and certainly by the late ninth century boom in literary output. If fyrd place-names can be taken alongside here names, then the period at which the names were coined might have coincided with the period at which fyrd began to develop its attested Old English meaning 'army', which it certainly had by the late ninth century, but when such a use was still relatively restricted. This would explain why fyrd names are much less common than those with the specific here. It is important to reiterate that fyrd is used in English toponymy, including sometimes in appellative compounds, and almost certainly with the sense 'army'; but that its use in that context is rare. 105 Even if some fyrd place-names are given an entirely different explanation, it would not undermine the rest of the argument set out here; and the question would remain: why, once fyrd had come to mean 'military force', and when it was an active place-name element with that sense, was it not used to coin place-names in the same way as here was? The choice of here in such compounds, to the exclusion of fyrd, would probably have much the same implications for the dating of hereplace-names. On balance, a period around the later eighth and early ninth centuries might seem reasonable on linguistic grounds for the coining of most of these here names, the reservations already expressed notwithstanding.
gEogrApHicAL Distributions In discussing the location of hereplace-names, it is clear that there is meaningful coincidence with the boundaries of shires, perhaps even with underlying 'tribal' regions ( Fig. 6). Essex, named after the territory of the East Saxons, is ringed by three hereplaces, one each for the neighbouring shires of Suffolk, Cambridgeshire and Hertfordshire. Only its border with Middlesex is lacking a named site, a fact which might anyway reflect the complex political relationships that existed between these districts, particularly in the seventh and eighth centuries. 106 In that case the location of Harefield (Mx) close to the meeting of Hertfordshire, Buckinghamshire and Middlesex may be significant. Other sites give the impression of being centrally placed between groups sharing a political identity. Thetford (19) lies close to the boundary of the South and North Folk of the former kingdom of the East Angles; Harlow Hill (2), lies close to the River Tyne, and therefore near that of the Bernicians and Deirans. 107 Clearly this is not straightforward liminality, for location at these boundaries actually endows the sites with an element of centrality. They sit on the borders between shires or possibly semi-autonomous regions within larger kingdoms or overlordships. This kind of internal liminality is shared by the places of muster named in narrative sources; the (likely) locations of Ecgbrihtesstan and Iglea (Iley Oak), two mustering points on Alfred's march to Edington, were near to shire boundaries, 108 and Wilfaraesdun must have been close to Gatherley Moor, towards the border of Durham and the North Riding of Yorkshire. It is also echoed at a local level in the frequent positioning of hundred meetingplaces on parish boundaries. 109 Most of the correlations with boundaries fall on shires which only make their appearance as administrative districts in the tenth and eleventh centuries; their interpretation depends to some extent on our evaluation of the antiquity of the boundaries that underlie shire geography. 110 In this regard the absence of comparable sites from Wessex (where the documentary evidence for shire divisions is somewhat earlier) is very striking. South of the Thames, the here names discussed here are rare, and neither they nor fyrd names have close correlation with the convergence of such major boundaries. The impression is that the hereand related names represent the focal points either of a formal system imposed from above (i.e. under West Saxon jurisdiction), perhaps at the same time as the formal rearrangement of the northern shires; or of a midland, East Anglian and northern system not adopted in Wessex and perhaps greatly altered during the West Saxon expansion. In any event, we should acknowledge the probability that other here names once existed that have since been lost, but the location of four such sites (five if Harwich is included) at intervals along the East Saxon/Middle Saxon frontier -effectively the West Saxon frontier for much of the ninth century -is perhaps telling.
As part of this pattern, the association of several sites (3,10,13,15) with the junction of three neighbouring shires is also noteworthy, particularly in light of the Chronicle account of Alfred's muster at Egbert's Stone. Various locations have been considered for the site of Egbert's Stone, 111 and all of them cluster close to the intersection of the shire boundaries of Wiltshire, Somerset, and Dorset. The Chronicle details earlier troop movements involving the men of more than one shire, in 860 (Berkshire and Hampshire) and 845 (Somerset and Dorset), 112 and it is probable that these mobilizations similarly required a predetermined assembly place close to the boundaries of neighbouring shires. 113 This class of place-names may therefore commemorate a stratum of mustering sites superior to those associated with hundreds and wapentakes, whether they originally belonged to the same chronological phase or not. This would explain why they so seldom appear in hundredal nomenclature. We might think of them as a West Saxon innovation, later rolled out as a formal system into previously Mercian territories, or as the remnants of an earlier system of large-scale musters, found across Mercia, East Anglia and Northumbria, on top of which a more regular and localized system of assemblies was later superimposed. The adoption of a unified nomenclature for these sites outside Wessex may favour imposition from above, based on a West Saxon model.
Neither interpretation can satisfactorily explain the location of all hereplaces. A case could be made for Harville next to Wye (12) lying on the edge of a core eastern Kentish kingdom, but this had certainly expanded further westwards thereby accounting for Alfred's short (one night) camp at the meeting place. Furthermore, Peddie, Alfred Warrior King, p. 135, argues that inter-shire rallying points would need to have been easily accessible and close to shire boundaries, given that such assemblies would have significantly weakened local military capabilities.
Access to communications appears to be the most significant issue underlying the location of Egbert's Stone. Peddie further argues that proximity to shire boundaries is likely to have been an issue. He suggests that shire fyrds would have stayed within their own territories until the last minute, only moving on to Egbert's Stone on the final day. Alfred himself is likely to have avoided northern trackways, preferring the Hardway from Athelney. These arguments appear to undermine the claims of Alfred's Tower and Coombe Street as the site of Egbert's Stone, as both are located too far west, and on the boundaries of shires not supplying militia for the fyrd. G. B. Grundy, 'The Ancient Highways and Tracks of Wiltshire, Berkshire and Hampshire and the Saxon Battlefields of Wiltshire', ArchJ 75 (1918), 69-194, and A. H. Burne, The Battlefields of England (London, 1996) support the Willoughby Hedge site, which does fit these criteria. This is the crossroads of Ridgeway, the Hardway, and the Grovely Ridgeway, but is quite an exposed location, and not directly accessible to troops moving from Hampshire, Wiltshire and Somerset. Instead Peddie favours the Kingston Deverill location on grounds of communications. There is here also a tradition of stones from nearby King's Court Hill, maintained to be the location of Alfred's court. 112  by the sixth century. 114 More likely, given its location on a major crossroads providing good links north-south and west-east across Kent, was that it served as an assembly place centrally located for easy access from the whole lathe of the Weowara 'people of Wye', or even for the entire shire. 115 The locations of Harlow in Edingley (Nt) (8), Harford (De) (16), and Hartford (Hu) (17), are similarly central to their respective territories. It may be the size of the gatherings that took place at these sites that led to their being singled out for association with a here or þēod, in contrast to other meeting-places at which only smaller groups gathered. The evidence for this group of place-names, such as it is, is consistent with a date just prior to the major West Saxon military reorganization of the late ninth and early tenth centuries, and this has important implications for their interpretation. If the names belong by and large to the middle Anglo-Saxon period, they may well represent a system of mustering that existed before the establishment of the hundredal system as it appears in the late tenth and eleventh centuries. In this there is convergence between the geographical location of here places at the margins of definable 'tribal' groups and the first mentions made during the ninth century to mobilizations of tactical groups named by their shire. The case for such sites existing around the territories of the East Saxons and the Hwicce has been outlined above, and it seems not unlikely that these represent the emergency mustering places of semi-autonomous districts; however, the lack of comparable here places in West Saxon and Mercian heartlands is worthy of note. The absence of these names from the major middle Anglo-Saxon kingdoms might suggest that there already existed in these areas localized systems of mustering foreshadowing that of the later hundredal system.

concLusions
In explaining place-names of the type discussed here -appellative compounds containing elements such as OE here -the tendency has been to look for a single unifying economic or topographical characteristic shared by all examples. 116  names: their association with large groups of people, their reference to obvious landmarks, and their partaking in the language and landscape of assembly. The distribution of these sites also makes clear they had a relation to large territorial units, defined -at least in part -by regional and 'tribal' affiliations. It may be significant that the Scandinavian hérað or hierat, a term derived from a cognate of OE here, is applied to administrative districts. 117 It is clear that places called here-hlāw could also become centres of administrative districts, but the primary sense of here 'army' suggests that a notion of the mustering of force existed from the start. As is shown above, these sites have a clear correlation with the boundaries of shires, territorial units that are most prominently associated with military forces in sources of the early ninth century. 118 This too hints at a military rather than a purely judicial role for sites named here-hlāw, -ford, and -feld. This model of military infrastructure supports the belief in the existence of a more general obligation affecting all freemen occupying a region, rather than the more specific infrastructures of aristocratic warfare.
Whilst the eighth and ninth centuries provide a likely context for the emergence of such sites, it is clear that regional levies were still expected to muster in specific conditions, as was the case in 878; the spatial correlation between herenames and later shire boundaries is again relevant. Only over the course of the tenth and eleventh centuries could this function have been superseded by a more localized network of hundred/burghal musters. The reforms of the late ninth century might be thought of as a move from use of militias to use of a fyrd not just for planned offensives, but also for reactive defensive campaigns. Such a change must have had a significant impact on the organization of mobilization. Under the latter system, the main centres of collection -for resources and for men, and later also for legal and administrative activity -were the strongholds or burhs. Just as importantly, armies were no longer only raised on a short-term basis, for a single engagement (or short series of engagements) or to respond to an immediate threat; instead, they could be raised with longer-term objectives in mind, a precaution against a more constant background menace of invasion, or the military arm in a long-running campaign of diplomacy, expansion and conquest. Furthermore, armies of the late ninth century and later were raised from a much wider area, so that the demand on manpower in any single district was much less, and the need for individual shire musters perhaps diminished.
The military reforms of the later ninth century must have had a significant impact on the composition of armies and their strategic performance, but it should not be assumed that earlier systems geared to the raising of armies were necessarily unwieldy. An army that mobilizes only to deal with present threats and disbands again after a short campaign must require a very effective system of musters, allowing the constitution of a fighting force in the shortest time possible. If the arrangements for mobilization became outmoded -and it is certainly not clear that they did -then it would have been because they were not geared to the new end-product of fyrd and burhs. They might have become inefficient within a drastically altered military system, but need not have done so within the system for which they originally evolved.
It is our contention that these changes can be traced in the organization of the landscape of mobilization, which suggests the coexistence of, on the one hand, a regularly ordered and probably centrally imposed system of assemblies that might easily have served also a military purpose; and on the other hand a number of major regional gathering sites whose place-names, location and landscape context are suggestive of use as musters for warriors. These two systems may have been contemporary, reflecting the hierarchical nature of mobilization; or they may have been chronologically distinct, a more complex system perhaps superseding an earlier territorial one to keep pace with the changing requirements of the military. Only by adopting a multidisciplinary landscape-based approach has it been possible to identify potential elements of these systems and to suggest subtle changes in the framework of mustering sites. Further multidisciplinary examination of these and similar sites may be able to add to our understanding.
In many ways, the foregoing discussion has only scratched the surface of what is a very deep and complex subject -that of military mobilization in Anglo-Saxon England. With a few notable exceptions, it is a subject that has been largely overlooked, at least in its landscape dynamics, by commentators on Anglo-Saxon warfare. Perhaps in the main this is due to the lack of written evidence pertaining to sites of muster. If, as seems likely, mobilization was based on the system that evolved into the network of hundred meetingplaces recorded in Domesday Book, then there is great potential to further our understanding of it, but the evidence of place-names suggests that many other important sites of military assembly may also have existed, and these too may reward further study. This article has demonstrated that a landscape-based approach can hope to reveal a great deal more about the systems of military mustering so crucial in valorizing military service and underpinning the military response of early medieval war leaders. 119 In this way, the heap of stones referred to in the wholly OScand name Harrogate might well have been the same heap named 'stone mound' by Old English speakers. A first element *haer would also seem to make better sense of the interchange between -arand -erin the early forms; the outcome of OE /ae/ being represented in ME orthography by both <a> and <e>. 127 The earliest -arform is very early to be a development from the -erof here-hlāw.
Three instances of Hare Law or Harelaw in Northumberland and Durham (4, 5 and 6) are taken by Mawer to contain OE hār or perhaps hara 'hare'. 128 Mawer took OE hār to mean 'boundary' in these instances, noting that the Pelton and Kirkharle examples were on parish boundaries, which may have influenced his interpretation of the etymology. That OE hār could mean 'boundary' was suggested by Hamper and Duignan, 129 who noted the frequent occurrence of the element in charter bounds and in place-names located near to boundaries. The actual meaning of OE hār, however, is 'grey, hoar', and this could be the meaning in most instances; certain features often used as boundary markers, such as stones, are also apt to appear grey (perhaps through being overgrown with lichen), so the frequent use of the term for boundary features may be no more than coincidence. 130 This being so, there is no need to interpret the Pelton and Kirkharle instances as hār names on grounds of their geographically liminal location, and there is nothing in the surviving forms that rules out derivation from herehlāw -though of course the forms are too late for certainty either way. If the names remained transparent during the ME period, forms in -orwould more strongly point to OE hār, but such forms are not known.
More difficulty arises from the early forms of Harlow Hill in Northumberland (2). The origin was taken by Mawer to be here-hlāw and Ekwall agreed, but thought the early forms remarkable. 131 Mills is also happy with this etymology. 132 It must be acknowledged, however, that the range of early -ir-, -yrand -erforms is problematic, and led Watts to suggest a compound OE higera, higre 'magpie' and hlāw. 133 While this may explain some of the earliest spellings, it does not account for the eventual -ar-forms, which start in the sixteenth century and could be indicative of original -er-. The development of ir<er does have parallels in northern England, and Mawer describes other examples, 134 but none of them precede the liquid r. Smith gives examples of ME er>ir and er>yr in West Yorkshire, 135 but the forms showing this development only go back to the fifteenth century. It is also worth noting that ME hēre was occasionally represented textually as hire. 136 In the final analysis, Harlow Hill in Northumberland defies secure explanation, but the range of forms is not out of keeping with OE herehlāw, even if the dates at which those forms occur are unusually early. It is included in this discussion, but with reservations.
Many other modern instances of Harlow lack forms earlier than the nineteenth century and their etymologies must remain in doubt. Harlow in Edingley (Nt) (8), and in Mayfield (St) (9), do at least have seventeenth-century spellings and are included in the table as possible instances of here-hlāw, but their forms are not early enough for any certainty.
These and subsequent place-names are in sequence according to their map number (Figs. 5 and 6).