Pilgrimage and procession: temporary gatherings and journeys between the tangible and intangible through the archaeology of South Asia

ABSTRACT South Asia is home to many of the world’s major religions, and over a century of archaeological enquiry has documented the sites associated with these traditions. Although textual scholarship, augmented by art historical and architectural studies of durable remains, has dominated interpretations, recent archaeological studies have begun to redress this balance by contextualizing monuments within their landscapes and engaging with the varied roles that such monuments played in the past. Referencing ethnographic analogy and archaeological visibility of ritual practices, alongside analysis of archaeological and textual evidence, the authors explore the issues faced when identifying and interpreting temporary gatherings at these sites in the past with reference to pilgrimage. Utilizing a broad-spectrum approach through varying time-periods and traditions, they advance potential ways of bridging the gap between intangible practices and tangible evidence, revealing the role of settlements, religious sites and landscapes as routeways and assembly points for pilgrimages and processions.


Introduction
South Asia, a geographically diverse region, is home to over a third of the world's population, several hundred languages and varied religious communities, including those from the major world religions of Buddhism, Hinduism, Islam, Jainism, Sikhism, Judaism and Christianity (Coningham and Young 2015, 3). The worship and veneration of these religions, and regional traditions, has led to the development and continued creation of monuments and sites, forming durable physical manifestations of social and religious loci in the past and present. Monuments from historic periods have traditionally been studied through an architectural and art historical lens (Ray 1994, 36;Trainor 1997;Trautmann and Sinopoli 2002), focused on the recording of structural plans, elevations and sculptural remains, linked to textual scholarship and inscribed and iconographic artefactual finds (e.g. Fergusson and Burgess [1880]; Brown [1956]; Korn [2007]; Mitra [1971]). Such approaches can provide a fairly static view of sites and a perception of monuments devoid of people and agents, even though modern practice illustrates that religious sites and settlements are the focus of varied and dynamic activities, especially during major festivals and pilgrimages. Frequently seasonal in nature, these events involve not only the agglomeration of large populations from wide areas and different communities, but also intangible practices in Referencing a single example, the Samai Mai shrine at Tilaurakot highlights the inherent difficulties in recognizing the impacts of pilgrimage and temporary gatherings in the archaeological record.
Further to archaeological visibility, there is the identification of variability in practice in the past. Tilaurakot again highlights this variability at a single site with local Hindu populations visiting the Samai Mai shrine and international Buddhist pilgrims venerating and meditating at conserved structures, including the site's Eastern Gateway where it is believed the Buddha left the city on his journey towards enlightenment. Rather than viewing sites as binary and linked to one particular tradition, they are often plural, and 'pilgrims of different faiths may visit the same pilgrimage sites, challenging the notion of clear divisions between the various religions of the region' (Branfoot 2006, 62). In India, sites like Champaner-Pavagadh in Gujarat, as well as Rajgir, in Bihar (Harding 2005, 48), are visited, and various parts of these sites are venerated, by those who identify as Muslim, Hindu, Buddhist and Jain. This is similar to Kataragama, a major pilgrimage site in Sri Lanka, where the shrine for the god Skanda is shared by both Buddhists and Hindus (Obeyesekere 1992). Even major international Buddhist centres, including Bodh Gaya, where the Buddha gained enlightenment, and also Lumbini in Nepal, the Buddha's birthplace, are not restricted to Buddhist pilgrims. Bodh Gaya, as an intersection of both Buddhist and Hindu pilgrimage, has been a contested and plural space since at least the nineteenth century (Geary, Sayers, and Amar 2012). Similarly, at Lumbini, in 2013, 50.90% of pilgrims to the site identified as Hindu, with the sculpture depicting Mayadevi giving birth to the Buddha also venerated as the local deity Rupa Devi (Coningham and Acharya 2013).
Recent landscape-based studies have placed large monumental religious and settlement complexes within their immediate settings, regional networks and pluralistic patronage and pilgrimage networks (e.g. Fogelin [2006]; Shaw [2007]; Hawkes [2009]; Willis [2009]; Coningham et al. [2013a]). By integrating data from multiple sources, including inscriptions, texts, sculpture, ethnographic observations and archaeological evidence, these studies are moving towards an approach that Willis has described as an 'archaeology of ritual' (Willis 2009, 3). In order to attempt to identify large temporary gatherings in the past, especially in relation to pilgrimage, we follow a similar approach, relating ethnographic observations of contemporary practice with textual sources and archaeological evidence from varied time-periods within a broad regional perspective ( Figure 2). Through these strands of evidence we will explore how large temporary gatherings may have transformed sites in the past. Finally, we focus on the pressures of temporary and transitory communities on major religious sites in the present and how influxes of large numbers of people can inadvertently damage the very monuments that they travel to venerate.

Textual evidence for pilgrimage and temporary gatherings
Afforded a primacy in scholarship within South Asia, textual sources have provided insights into the presence and practice of pilgrimage in the past. The earliest references to pilgrimage can be inferred from Brahmanical sources such as the Rig Veda (Guy 1991, 356). Much later fifthand sixthcentury CE Hindu Puranic literature records religious sites referred to as tirthas, or holy crossing places, and the locations of these pilgrimage destinations (Eck 1981). Written centuries after the origins of the sacred sites they reference, they do suggest the presence of locations of ritual foci within landscapes. In some instances they record specific lists of sacred sites as well as the practices that were carried out at these locations, including the veneration of shrines, ritually purifying bathing and the performance of rites of post-mortuary ancestor worship (shraddha) (Entwistle 1990, 7).
Epigraphic sources can also provide evidence for pilgrimage in the past. For instance, the corpus of third-century BCE inscriptions attributed to the Mauryan Emperor Asoka (r. 272-235 BCE), and established across South Asia (Falk 2006), not only provides insights into the sociopolitical organization and the territorial extent of this major regional power (Smith 2005), but also evidence of patronage. A prime example is the text engraved on the pillar at Lumbini, erected in commemoration of Asoka's personal pilgrimage to the site, which states: [W]hen King Piyadasi (Asoka), the beloved of the gods, was consecrated for this 20th regnal year he came in person and paid reverence. Because the Buddha, Shakyamuni, was born at this place, he had a stone railing made and stone pillar erected. Because the Lord (of the world) was born at this place, he exempted the village of Lumbini from taxes and granted it the eight shares. (Falk 2006, 180)  A later act of pilgrimage to Lumbini is also recorded on this pillar, with a fourteenth-century CE inscription added by King Ripu Malla of an independent kingdom in the hill region of Nepal. This indicates the continued importance of pilgrimage to Lumbini, even during a time period when it is traditionally assumed that this region was under low-level occupation outside the realms of a centralized authority (Bidari 2013, 54).
The travel accounts of Chinese pilgrim monks, including Faxian (337-422 CE) and Xuanzang (602-664 CE), also provide further evidence and details of pilgrimage and pilgrim routes throughout Asia. Listing pilgrimage destinations and corroborating the presence of monuments at major monumental sites, such as the Asokan pillar at Lumbini (Deeg 2003, 50-5), these accounts also record religious festivals and practices as well as settlements, communities and landscapes encountered during these journeys (Bidari 2013). For instance, Faxian recorded festivals organized for the display, procession and veneration of relics and deities, such as at the former Mauryan capital of Pataliputra (Ray 1994, 38), as well as a three-month long festival in honour of the Buddha's Tooth Relic at Anuradhapura in Sri Lanka (Nissan 1988, 269). Similar practices at Anuradhapura are referenced within Sri Lanka's Pali Chronicles, including the Mahavamsa, which describes festivals from the arrival of Buddhism in the third century BCE onwards, including the replastering of the Ruvanvelisaya stupa and the watering of the Bodhi tree, as well as the consecration of monuments and enshrining of relics in the presence of large gatherings (Geiger 1912).
Although providing evidence of intangible practices and large gatherings in the past, textual sources often relate to elites and an idealized version of events (Schopen 1997), and may not fully reflect the experiences of all strata of society and varied communities who undertook spiritual journeys or participated within festivals and gatherings. However, it is clear that religious sites were major focal points for bringing large populations together, and archaeological evidence suggests that this was both through their role as pilgrimage centres, but also as sites linked to networks of trade and exchange.

Religious sites, markets and temporary gatherings
In the nineteenth century, partly on his observations of contemporary pilgrims, Sir Alexander Cunningham suggested that hundreds of thousands of small baked and unbaked clay stupas, found during renovations at Bodh Gaya, were offerings left by pilgrims (Cunningham 1892, 47). Similarly, models of temples have been interpreted as pilgrim souvenirs and the distribution of find spots of models of Bodh Gaya's Mahabodhi temple have been inferred as evidence of pilgrimage networks across South Asia (Guy 1991, 364). Evidence of such networks is clearer from the presence of ceramics and materials relating to production and exchange linked to international trade.
For many regions and religious traditions, it has been noted that the landscapes and routeways through which pilgrims would have travelled would have been 'corridors charged with economic and social activity of a multitude of kinds' (Stopford 1994, 60). In South Asia, it has been suggested that many ritual and religious sites emerged or were located along trade routes and important crossing points (Barnes 1995;Chakrabarti 1995, 199;Heitzman 1984, 124, Hawkes 2009Rees 2010). It has been postulated that some of these locations were of pre-existing ritual, and perhaps economic, significance, which were targeted by emerging religious movements, such as Buddhism, for the establishment of sites and influence in the landscape (Hawkes 2009, 156-7;Shaw 2007, 55, 127). In the Early Historic Period (c.500 BCE-320 CE), epigraphic evidence suggests that merchants were pivotal to the establishment of these foundations through donations and support (Ray 1994, 38) and during this time period, Buddhist monasteries have been interpreted as consumers as well as traders and an important component within the economies of polities (Kosambi 1955;Morrison 1995, 205). Whether these religious sites were the catalysts for, or beneficiaries of, trade and whether pre-existing exchange networks and routeways were coopted or were established through these religious foundations, settlement surveys suggest that these sites were the focus of social organization and centres of redistribution, production, exchange and in some instances controlled hydraulic infrastructure for agriculture (Bandaranayake 1994, 16;Coningham et al. 2013a;Fogelin 2006;Shaw 2007). Such varied roles placed religious institutions at the centre of communities, but also led to them becoming hubs for contacts from further afield.
At Abhayagiri and Jetavana at Anuradhapura, and at monastic sites in the city's hinterland, the recovery of sherds of Northern Black Polished Ware and Rouletted ware in Early Historic period deposits indicate links to northern India as well as engagement with Indian Ocean trade networks (Bouzek 1993, 83;Prickett-Fernando 1990, 81;Coningham et al. 2013a, 463). A similar pattern has been identified throughout South Asia and at Thotlakonda, investigations have uncovered evidence of local exchange links, through finds of regional ceramic types, as well as contact within wider international networks through the discovery of Rouletted Ware and Roman and Satavahana coins (Fogelin 2006, 92, 154). Not just recipients and receptacles of trade, religious sites also engaged in craft production, with evidence of metalworking identified at the central monasteries of Anuradhapura, as well as within the hinterland (Coningham et al. 2013a;Hettiaratchi 1994, 67-8;Bouzek 1993, 107). Further ranges of craft activity were recorded at the core, including tile glazing workshops at Abhayagiri (Bouzek 1993, 13) and glass production at Jetavana (Ratnayake 1984, 199). Such production is not limited to Buddhist sites and recent excavations at the medieval Hindu Temple of Shiva Devale No. 2 at Polonnaruva in Sri Lanka has uncovered evidence of large quantities of iron slag as well as a gold mould, suggestive of production aimed at local and potentially wider regional markets. This suggests that religious institutions were central within networks of production, services and exchange. In prime locations on established routeways, this central role would have led to these sites acting as market centres for both permanent and temporary populations from a wide range of geographical distances. These markets alongside the religious roles of these sites would provide for general needs throughout the year, but also for short-term population peaks created by specific events and festivals.
For instance, monumental stupas at the large Buddhist monasteries within the city of Anuradhapura have acted as gathering points for large religious festivals for local communities as well as national and international pilgrims. Such occasions include the Jasmine Flower Festival, where thousands of family groups from the hinterland, and further afield, bring offerings including rice, coconuts and money to central Anuradhapura (Coningham et al. 2013a, 463). The estimated capacities for the courtyards and terraces that surround Anuradhapura's major stupas of Abhayagiri, Jetavana and Ruvanvelisaya suggest gatherings of between 9300 (at 3.6 per square metre per person) and 75,000 people (at 0.46 per square metre per person) were possible, whilst smaller stupa terraces, such as at Mirisivati could host between 1100 and 9200 people. It is noted that these estimates do not account for multiple visits throughout the duration of a festival that may last a number of days. Furthermore, pilgrims may also reside in the open areas surrounding monastic institutions suggesting that the sizes of these temporary gatherings may have been much higher (Coningham et al. 2013a, 470). This is a proposition supported by contemporary observations and records of large pilgrimages to Anuradhapura in the nineteenth and early twentieth centuries, with 25,000 people recorded at a festival in 1897 and an attendance of 200,000 people at the enshrining of relics at Ruvanvelisaya stupa in 1932 (Nissan 1988, 256).
Not only do these terraces provide large spaces for the agglomeration of vast numbers of people for temporary gatherings (Figure 3), the construction of these spaces would also have involved the coming together of large communities in the past. Whilst the construction of monuments is often attributed to individual elite and royal patrons (Thapar 1992, 19), epigraphic evidence, at sites such as Bharhut, Sanchi, Pauni and Anuradhapura, illustrates that many monuments were created and constructed through a process of individual and collective donations (Dehejia 1992, 35;Fogelin 2003, 140;Sawant 2011, 109;Davis et al. 2013, 448). This is evident today: during recent renovation at Anuradhapura's Ruvanvelisaya stupa, the site manager encouraged volunteers to help with the restoration of its large courtyard sand terrace. A total of 16,668 individuals from 324 settlements came to the site to volunteer (Gunawardhana 2017, 10), and illustrates that the meritorious act of aiding the renovation and construction of a religious monument can also lead to large temporary populations.

Impermanence and pilgrimage
Once constructed, successive groups of pilgrims and worshippers will then interact with the monuments. Within archaeological interpretations, sight-line analyses of surviving ground plans have been undertaken to identify areas of communal and individual worship at Buddhist monastic sites. These have particularly focused on the role of site layout in maintaining views of stupas as a way of gaining merit through the act of 'seeing' monuments linked to venerated objects (Shaw 2007, 142-3;Fogelin 2011, 205-6). These studies have provided new insights into how monuments that had previously been viewed as static, and without human engagement, may have been spatially positioned and interacted with. However, monuments are not always permanent or constructed from durable materials. During festivals, temporary structures, sometimes on a large scale, can be erected, including stalls set up to cater for the needs of pilgrims selling food, drinks and materials for worship and veneration. They can also include temporary spaces created for devotional practices. Therefore, whilst sight-line analysis provides new interpretations of how space and movement may have been created and managed, it does not account for the potential of temporary structures to obscure or reveal elements of a site. This can be demonstrated by observations made during fieldwork at Polonnaruva's Shiva Devale No. 2, in Sri Lanka, in 2015 and 2016. The shrine within the main sanctum of the stone temple and the Nandi sculptures within its precinct were venerated by devotees throughout the week . Rituals were also undertaken at a tree, outside the temple precinct, that had no durable or visible signs that it was part of the ritual structure of the site apart from its use by devotees. When a larger puja was held, a temporary shelter was erected in front of the sanctum of the temple, constructed of scaffolding and tin sheets. Extending the footprint of the temple, the erection of this temporary canopy completely changed the routeways around the temple, and how space was used by those visiting the site. During this gathering, new areas of the site became the focus for further activities, including the preparation and cooking of offerings. Within a few days the shelter was dismantled and the temple activities continued as before.
Sometimes completely new temples are erected during festivals. During Shivaratri at Pashupatinath temple, in Nepal, a large temporary temple, constructed of bamboo and plastic, is built within the monument complex to accommodate the influx of pilgrims who wish to attend temples and the temporary markets set up during the festival (Figure 4). Further examples of such temporary structures are found throughout the year at Pashupatinath, including a Shivalinga constructed from fabric, which was built during observances for Bala Chaturdashi, when those who have died within the past year are remembered ( Figure 5). Erected for short time periods coinciding with festivals and events, these vestiges are then dismantled. These non-permanent and seasonal structures and monuments, and their uses by people, are potentially missing from archaeological observations, particularly those that traditionally focus on durable, and aboveground, architectural remains. Whilst discussing Pashupatinath, it must also be noted that it is not only pilgrims who travel through the landscape to visit and gather at religious sitesso too do religious practitioners. Pilgrimage locales can become gathering places for usually dispersed communities, such as wandering ascetics and networks of pilgrimage, often linked to sacred calendars, enable travelling renouncers to periodically gather at predetermined locations, facilitating ritual interaction and economic support systems for renouncers and their followers (Hausner 2007, 105). At Pashupatinath, Sadhus throughout South Asia congregate throughout the year, most spectacularly at Shivaratri, where in 2016 over 5000 Sadhus were recorded amongst the 1.5 million pilgrims who visited in a two-week period (EKantipur 2016). At the end of these festivals, physical evidence of these large gatherings almost vanishes with temporary structures dismantled and pilgrims and Sadhus dispersing from the site back to their communities or wandering lifestyles.

Pilgrimage across the landscape
The impact of transient populations is not limited to pilgrimage destinations, but also affects the landscapes they move through (Mack 2004, 64), including the creation of infrastructure intentionally and also as a by-product of pilgrimage. Whilst the destinations of pilgrimage have been studied, there is a potential to identify and understand the places between major destinations where large populations would have temporarily stayed during their journeys. For the major Islamic pilgrimage of the Hajj, features along established pilgrim routeways, such as caravansarais, mosques, forts, cemeteries, cisterns and water systems, as well as the organic growth of settlement around such infrastructure, are being identified and interpreted archaeologically (e.g. Peterson [1994]; Petersen [2012]; Blair and Ulrich [2013]). Such infrastructure is now being identified in South Asia, and recent fieldwork in Nepal, near Lumbini and Tilaurakot, has identified sites that may have formed waypoints on routeways through the landscape.
Karma and Dohani are located between Lumbini and Tilaurakot and these sites are spaced at roughly 12-kilometre intervals from each other. Both Karma and Dohani comprise squared earthen enclosures, measuring 50 metres by 50 metres, with a rounded stirrup bastion on each corner, similar in appearance to Kushan period (c.30-225 CE) sites such as Sirsukh in the Taxila Valley, Pakistan (Marshall 1951). The almost identical design of Dohani and Karma, and their spacing within the landscape, indicates centralization and the creation and management of waypoints. Though both sites have not yet been scientifically dated, recent excavations conducted the Department of Archaeology (Government of Nepal) at Dohani, identified a deep occupation sequence, whereas auger coring across Karma revealed no evidence of earlier cultural material and habitation (Kunwar et al. 2018). This suggests that existing settlements, such as Dohani, were incorporated as waypoints, and sites like Karma were created to form new stopping points, or to formalize temporary camps on routeways that had little prior archaeological visibility. At Dohani, magnetometer survey has identified that the majority of its interior did not contain evidence of durable structures, and it has been suggested that similar seemingly empty sites may have provided bounded spaces for temporary camps used by mobile populations (Coningham and Manuel 2009, 239). Whether for trade, landscape administration, pilgrimage or multiple uses, these sites suggest control of routeways and the formalizing of locations where populations could gather for short periods of time. Temple and monastic sites across the landscape may also have been used for shelter on journeys and informal/temporary shelters erected at these sites and along pilgrimage routes. Such temporary camps are seen during the two-week-long Pada Yatra, or foot pilgrimage, of Hindu Tamil communities from Jaffna and the east coast of Sri Lanka who walk up to 300 kilometres to reach the pilgrimage site of Kataragama in southeast Sri Lanka. During each daily stage pilgrims set up daily camps at temples along the eastern coast road, which vanish once the pilgrims leave and the pilgrimage season ends (McGilvray 2010, 356). The evidence outlined here earlier illustrates that religious sites can form locales for large gatherings, seasonally or throughout the year for communities and individuals on local and international scales. This population increase, and increased need for services, amenities and goods, has led to these sites, and the corridors in between, becoming major market centres and networks, throughout the year or during annual pilgrimage calendars. This contrasts with nineteenth/twentieth-century notions that religious sites, particularly monastic establishments, were places of solitude and isolation, cut adrift from worldly concerns (Fogelin 2006, 52). Such gatherings are not restricted to religious sites and the routeways between, but can also be explored within urban environments.

Gatherings and processions within urban forms
Urban forms attract disparate populations and large agglomerations of people, sometimes for seasonal or long-term durations, fulfilling roles as economic and potentially emotional, anchors for communities (Smith 2006, 130). More than just economic hubs, settlements from various time periods throughout South Asia have been analysed in relation to their symbolic layouts and attributes, cosmologically organized and designed as ritual and ceremonial centres (e.g. Wheatley [1971, 225, 481]; Coningham [2000]).
It has been suggested that routeways, within and around settlements, may have been manipulated to control access as well as movement, and therefore experience, through the creation of sight lines and access points (Mack 2004, 65). Rather than viewing the locations of structures and architecture within a settlement as static, movement through a city would have been a dynamic aspect of past populations, potentially providing an influx of population through, or to, specific locations. Recognized at several South Indian Temple towns, such control of movement has been identified at urban forms, such as thirteenth to sixteenth-century CE Vijayanagara (Mack 2004;Sinopoli 2004, 263), but most explicitly at Warangal, founded as the Kakatiya dynasty capital between 1199-1294 CE. The circular city is defined by three concentric ramparts broken by passageways and gateways, controlling movement by forcing a right turn into the city. From the centre, a radial system of eight roads fans out and with the imposed right turn creates three superimposed svastikas, an auspicious mark and solar emblem, as well as evoking mandalas, miniature reproductions of the universe (Michell 1992, 15-16). The layout of the city would mean that those entering and moving through its roadways would contribute to the creation of the auspicious svastika as well as mandala imagery, placing the ruler at Warangal at the centre of this microcosm of the universe (ibid, 16). Therefore, the movement and gathering of people, and the interactions of people with architecture, created meaning. However, how far these routeways were followed or understood is questionable, as, firstly, these plans do not account for the potential of temporary structures as mentioned previously. Secondly, in modern urban areas socalled 'desire lines' have been recorded, whereby inhabitants transgress imposed routeways, creating new routes and shortcuts where no formal paths have been provided (Moore 2017).
With this caveat highlighted, for the most part in the past and the present, civic and domestic architecture, as well as open spaces, guide the movements of inhabitants and visitors through urban sites (Smith 2014, 308). Open spaces within settlements have been frequently neglected and deemed unimportant by archaeologists due to a lack of structural remains to investigate. However, these seemingly empty spaces often provide the settings for social interactionsused for political gatherings, festivals and markets ranging in time-scales from daily, monthly and seasonal usage to unprecedented one-off events (Smith 2008, 218-21). The same spaces within a city that at one time might be used for a yearly religious gathering and procession may house a regular market and also be the location that a populace utilizes to process harvests. Excavation and geophysical survey at the Early Historic city site of Sisupalgarh, in eastern India, has begun to piece together the layout of this vast site. At the centre, several monolithic stone pillars have been identified, forming a defined monumental focus bounded by an empty space defined by a wall. It has been suggested that the creation of this space, as well as the sight lines towards it may 'have provided a space for performance and communal activities framed by the backdrop of this very imposing structure' (Smith 2008, 227).
The urban spaces of the Kathmandu Valley highlight how temporary gatherings can transform these open spaces. Settlements are not only the location of buildings but are also the abode of people, spirits and gods. The periodic performance of rituals, by communities within the city, maintains the city as an image of the cosmosand the spatial units and levels of social hierarchy are realized in repeated symbolic action (Gutschow 1993, 182;Levy 1990, 197). Ethnographic observations have recorded how urban communities partially conceptualize space and negotiate social relations through the movement of people during festivals and processions. In Bhaktapur, there are 24 districts within the entire settlement divided into two opposing halves, with the upper town in the east and lower town in the west (Gutschow 1993, 174;Levy 1990, 168, 183). This division is expressed during the Bisket-jatra New Year festival through ritually organized antagonism. A ceremonial chariot, reassembled each year for this purpose, is pulled by each half of the town in both directions, creating a ritualized struggle until it reaches the dividing line between each town half on New Year's Eve (Gutschow 1993, 174;Levy 1990, 170). Although several smaller festivals, limited to specific communities and deities occur throughout the year, Bisket-jatra involves almost the entire population of Bhaktapur with an estimated 20,000-30,000 people filtering into the main routeway and open spaces of the town (Gutschow 1993, 182).
As in Bhaktapur, festivals take place within the Durbar, or palace, squares of the old cities of the Kathmandu Valley throughout the year, transforming what appear on plans as static spaces into dynamic centres. Whilst buildings do anchor urban space and symbolic meaning, temporary gatherings can provide memories of spaces that have been lost. Processions and gatherings can reaffirm structures and spaces no longer visible. In Kathmandu Durbar square, around Hanuman Dhoka palace, families have been recorded following a processional route that follows the now disappeared wall's alignment, retaining civic memory of a now lost spatial marker (Slusser 1982, 93).

Temporary gatherings and the protection of heritage in the present
Religious sites and settlements represent more than static collections of buildings and are forums for dynamic interactions between tangible remains and intangible practices. The integration of textual, archaeological and ethnographic observations provides evidence of how people in the past may have utilized, changed and developed sites during temporary gatherings, particularly in relation to pilgrimage. Whether through the development of sites as markets, the erection of temporary structures or processions and movement through built architecture providing social meanings, sites and settlements have been transformed by the appearance and presence of large and mobile transient populations. Though the influx of large numbers of people at religious sites and settlements for temporary activities has a great antiquity, such practices continue in the present. Large numbers of visitors create challenges for site managers on how best to protect monuments, whilst also promoting sites and providing facilities and infrastructure that is not of detriment to finite archaeological remains. This is a pressing issue, as a report of 2012 estimated that 980 million tourist journeys were undertaken globally in 2011, including pilgrimage as one of the fastest growing motivations for travel, accounting for around 600 million of these journeys (SEDF 2012).
Lumbini provides a case in point, and has developed dramatically since its nomination as a UNESCO World Heritage Site in 1997. A large concentration of hotels and restaurants have been established around the site's perimeter since the development of the site's Masterplan and annual visitor numbers have increased from 17,000 to 800,000 between 2000 and 2011. The Asia Development Bank has estimated that these numbers will exceed 2,000,000 by 2020 and the creation of an international airport at nearby Bhairahawa will lead to further increased visitor numbers and associated infrastructure, threatening the heritage of the region through unchecked development (Coningham et al. 2013b(Coningham et al. , 1107. Archaeological research at Lumbini and in the wider region has both recognized the early origins of the site as a pilgrimage destination and is identifying areas for protection from development (Coningham et al. 2013b(Coningham et al. , 1119(Coningham et al. , 2017. This is aiding the guidance of the placement of infrastructure to serve pilgrims and tourists in areas of little archaeological sensitivity. Furthermore, within protected areas, new non-intrusive and reversible walkways are being installed, which allow visitors to walk over sensitive archaeology, negating damage from increased footfall ( Figure 6). Such interventions can aid the protection of heritage, whilst also facilitating continued sustainable growth and the economic benefits of large temporary populations at sites with a long tradition of temporary gatherings and pilgrimage in South Asia.

Notes on contributors
Christopher Davis is a postdoctoral research associate as part of the UNECO Chair on Archaeological Ethics and Practice in Cultural Heritage at Durham University. With a geographical focus of South Asia, his research interests include the social and economic roles of religious institutions in past societies, as well as the role of archaeological research for heritage management in the present and for the future.
Robin Coningham holds the 2014 UNESCO Chair on Archaeological Ethics and Practice in Cultural Heritage at Durham University. His research interests include urbanization, the development of Indian Ocean trade and the archaeology of early Buddhism. Undertaking fieldwork throughout South Asia, he currently co-directs archaeological explorations within the Natal landscape of the Buddha, including the site of Tilaurakot-Kapilavastu, as well as post-earthquake archaeological investigations within the Kathmandu Valley.