SETTLEMENT 5.1 Scope: settlement was an environmental negotiation A settlement map is not simply a map of where people preferred to live. In the Himalayan foreland and Middle Ganga plain, the ground itself was repeatedly remade by rivers, sediment, monsoon floods, groundwater, erosion and local tectonic movement. The same processes that threatened houses could renew fields, create fisheries, open transport corridors or bury older occupations. This chapter therefore treats settlement as a negotiation among several changing variables rather than as a one-time choice of a “good” site. The purpose is not to make environment a deterministic cause of culture. Human knowledge, labour, storage, social cooperation, route networks and technology all affected how communities used a given surface. 5.2 A flat plain contains many different surfaces At regional scale north Bihar and the adjoining Nepal Tarai appear extremely flat, but that appearance conceals differences that mattered greatly at household scale. Natural levees stand slightly above adjoining backswamps; abandoned channels become linear depressions; crevasse splays and bars create locally raised sandy ground; older alluvial surfaces may drain differently from younger floodplain deposits; and fan surfaces grade southward from the Himalayan front. Sinha’s work on the Gandak–Kosi interfan and later studies of the Gangetic plains emphasize this geomorphic diversity. A height difference of only a metre or two can change the duration of waterlogging, the accessibility of a well, the suitability of a crop, and the probability that a house floor remains dry during ordinary floods (Sinha 1995, 1996; Sinha et al. 2005). 5.3 Rivers of north Bihar do not all behave alike Sinha and Friend classified rivers of the northern Bihar plains by source area—mountain-fed, foothills- fed, plains-fed and mixed-fed—and showed that their hydrology and sediment behaviour differ (Sinha and Friend 1994). A large Himalayan river brings a different combination of discharge, sediment calibre and seasonal variability from a smaller plains-fed stream. This matters for settlement history because “living by a river” could mean living beside a braided trunk channel, an anabranching foothills-fed river, a smaller meandering stream, or an abandoned channel that retained water after the active course moved elsewhere. The economic possibilities and risks were therefore not uniform across the floodplain. 5.4 Megafans and interfan plains created different settlement problems The Gandak and Kosi have built enormous sedimentary bodies commonly described as megafans, while the terrain between them includes interfan tracts crossed by rivers such as the Bagmati and Burhi Gandak. Megafans are not static cones. Their surfaces contain active and abandoned channels, splays, flood basins and soils of different ages. Interfan plains likewise record channel avulsion, overbank sedimentation and waterlogging. Settlement history must therefore be reconstructed at the scale of individual landforms and local sequences, not by assuming that one river model describes the whole of Mithila and Vajji. HISTORY OF MITHILA, VAJJI & ANGA — VOLUME II Figure 16 — Landscape units and settlement opportunities across the Himalayan foreland–alluvial plain. The transect is schematic and not drawn to scale. 5.5 The Gandak megafan: long-term shifting changed soil and settlement surfaces Mohindra, Parkash and Prasad used geomorphology and soil development to argue that the Gandak channel shifted substantially eastward during the later Holocene, producing a sequence of geomorphic surfaces and soils of different ages (Mohindra, Parkash and Prasad 1992). Their reconstruction is important here not because every number must be transferred directly into archaeological chronology, but because it demonstrates that the apparent plain contains a history. Older surfaces may have had longer periods for soil development and settlement stability; younger surfaces may contain fresher sediment, abandoned channels and more recent disturbance. Archaeological survey that ignores landform age risks comparing areas with very different probabilities of preserving or exposing sites. 5.6 The Kosi megafan: do not repeat the simple westward-sweep story The Kosi is often described as having migrated steadily westward across more than a hundred kilometres before being embanked. Chakraborty and colleagues re-examined historical maps and near-surface sediments and challenged that simple unidirectional narrative. They argued that the historical channel record is better understood as oscillation and random nodal avulsion across different fan lobes rather than a continuous sweep (Chakraborty et al. 2010). For settlement history the methodological lesson is decisive: abandoned channels cannot automatically be placed in a neat west-to-east or east-to-west chronological ladder. Each palaeochannel and occupation surface requires independent dating and stratigraphic control. 5.7 The Bagmati shows how often a river can reorganise an inhabited plain The Bagmati provides a well-studied example of a hyperavulsive, anabranching river. Jain and Sinha reconstructed repeated major and minor avulsions over the historical period and related them to overbank flooding, high wash load, channel aggradation, local slope and tectonic influences (Jain and Sinha 2003, 2004, 2005). The lesson extends beyond modern flood hazard. A settlement beside one branch might later stand beside an abandoned channel; a route crossing could become unusable; a former wet depression might receive sediment and become cultivable; another tract could become waterlogged. Settlement continuity at a named locality therefore need not imply continuity of the same river relation. 5.8 Aggradation raises channels as well as land Overbank deposition can slowly raise floodplain surfaces, but sediment accumulation within and beside a channel can also reduce the relative depth of the channel and make avulsion more likely. Sinha’s work on GAJENDRA THAKUR northern Bihar emphasizes the combination of high sediment loads, rapid aggradation and low regional gradients (Sinha 1998). Communities living in such settings could benefit from fine silt deposited on fields while simultaneously facing increasing flood or channel-shift risk. The same sediment is thus both productive resource and geomorphic hazard. Figure 17 — Four fluvial processes that repeatedly reorganise settlement landscapes: lateral migration, cutoff, aggradation and avulsion. 5.9 Natural levees offered advantages, not immunity Natural levees form where coarser suspended sediment is dropped close to a channel during overbank floods, leaving slightly raised strips beside the river. Such surfaces can drain faster than adjacent flood basins and may therefore attract paths, fields and habitation. Yet a levee is created by flooding and remains part of the floodplain. It can be cut by bank erosion, breached by crevasse channels, or bypassed after avulsion. A raised settlement surface should therefore be understood as a risk-management choice within a floodplain, not as withdrawal from the floodplain. 5.10 Oxbows, abandoned channels and backswamps were productive spaces The low areas between active channels preserve water, fish, aquatic plants, reeds, clay and grazing opportunities. Chapters 3 and 4 showed that early agrarian economies did not rely on fields alone. Wetlands extended that mixed economy into the settlement landscape. An abandoned channel could serve as a water source or fishery long after it ceased to carry the main river. Conversely, prolonged waterlogging could restrict house construction and some crops. The economic value of a wetland therefore depended on season, access, technology and household needs. 5.11 Flood was not one event category Historical language often collapses several processes into “flood.” For settlement analysis it is useful to separate at least overbank inundation, prolonged waterlogging, bank erosion and avulsion. Ordinary inundation may leave fine sediment without destroying a settlement; bank erosion can physically remove the archaeological record; waterlogging can make an otherwise undamaged locality temporarily unusable; avulsion can transfer the main river into an entirely new corridor. These distinctions prevent modern disaster vocabulary from being imposed indiscriminately on archaeological deposits. HISTORY OF MITHILA, VAJJI & ANGA — VOLUME II 5.12 Sediment could improve a field and bury a house Fresh alluvium may replenish nutrients or create new cultivable surfaces, but deep sand deposition can also render fields temporarily less productive. Fine overbank silt, coarse crevasse deposits and channel sands have different effects. Archaeologically, the same deposit may seal a floor and preserve it from later disturbance. Thus a sediment layer is not merely “flood damage”: it can be a record of the event that ended one occupation surface and created the substrate for another. 5.13 Soil age matters to archaeological expectations The Gandak megafan study by Mohindra and colleagues linked geomorphic surfaces to differences in soil development. The broader implication is that older stable surfaces have had more time for pedogenesis, cultivation, erosion and repeated occupation, whereas young surfaces may bury older land beneath metres of alluvium. Survey intensity should therefore be adjusted to landform history. A dense scatter on one old surface and few sites on a recently aggraded tract cannot be read directly as a prehistoric population map. 5.14 Tectonics belongs in the settlement story, but not as a universal explanation The Himalayan foreland is tectonically active, and Jain and Sinha documented channel anomalies in the Bagmati basin associated with subsurface structural controls and changes in local slope (Jain and Sinha 2005). Earthquakes can alter gradients, liquefy sediment, damage settlements and modify drainage, but not every channel shift should be attributed to tectonics. Hydrology, sedimentation and local topography also matter. The responsible historical formulation is therefore multi-causal: tectonic deformation can change the conditions under which fluvial processes operate. 5.15 Monsoon variability altered both opportunity and risk The annual monsoon concentrated water and sediment delivery into a seasonal cycle. Stronger or differently distributed rainfall could increase flood frequency, bank erosion or channel activity; weaker phases could reduce surface water while exposing new bars and floodplain surfaces. Long-term alluvial sequences in the Ganga plain preserve responses to changing monsoon conditions, but settlement history should not assume that every cultural transition was climate-driven. Climate modifies a field of possibilities that communities respond to through cropping, storage, mobility, exchange and site choice. Figure 18 — Archaeological visibility is produced by geomorphology as well as human occupation. GAJENDRA THAKUR 5.16 Archaeological visibility is a geomorphic process One of the most important consequences of an active alluvial landscape is methodological. A village may disappear from the modern surface because it was buried beneath later sediment rather than because the area was abandoned. Elsewhere, bank cutting or gully erosion may expose artefacts and make a site unusually visible. Corvinus’s work in Nepal repeatedly benefited from erosion that exposed archaeological material, while the deep alluvium of the Gangetic plain can conceal occupations. The archaeological map therefore records a history of discovery conditions as well as a history of settlement (Corvinus 2007; Darnal 2016). 5.17 Burial can preserve what surface exposure destroys Rapid burial may protect floors, hearths, charred seeds or bones from later ploughing and weathering. The same process can place them beyond the reach of ordinary surface survey. Conversely, an eroding mound may yield abundant artefacts while losing stratigraphic context. This is why settlement counts must be accompanied by information about geomorphic setting, sediment depth and discovery method. “More sites” may partly mean “more visible sites.” 5.18 River-cut mounds and terraces deserve different expectations Chirand, Chechar, Panda and Oriup occupy different local settings, and their archaeological sequences cannot be reduced to one settlement template. A mound near a major river can preserve repeated occupations because people returned to or rebuilt on a relatively favourable elevation. A terrace-edge workshop such as Patu in eastern Nepal represents another relationship to topography and raw material. Comparing these places is useful only when the differences in landform, function and chronology remain explicit. 5.19 Sedentary settlement did not mean a fixed productive territory A house can remain in one village while fields shift among available plots, cattle move to seasonal grazing, fishers use changing channels, fuel collectors visit forest margins and traders travel to other settlements. The mixed economies reconstructed in Chapters 3 and 4 therefore imply routine mobility within sedentary life. Archaeological categories such as “village” and “site” can obscure this wider taskscape if they are treated as self-contained economic units. 5.20 Seasonal mobility reduced risk Floodplain households could respond to seasonal changes by moving activities rather than residences. Wet-season fishing, dry-season grazing on exposed surfaces, collection of reeds or fuel, and cultivation in differently drained plots spread risk across the landscape. The specific calendar varied locally, and the evidence does not permit a single prehistoric schedule. What can be defended is the structural point: diversified subsistence made spatial flexibility economically valuable. HISTORY OF MITHILA, VAJJI & ANGA — VOLUME II Figure 19 — Mobility without a false sedentary/nomadic binary: daily movement, seasonal resource use, exchange, temporary relocation and permanent site shift form a continuum. 5.21 Paths and crossings moved with the rivers Rivers were transport corridors but also obstacles. Fords, ferries and shallow crossings could shift as channels migrated or deepened. A settlement’s advantage as a crossing point was therefore historically contingent. Routes often followed levees or other better-drained ground, while wetlands could redirect travel. These changing connections help explain why settlement importance cannot be inferred solely from straight-line distance between archaeological sites. 5.22 Exchange linked contrasting ecological zones The alluvial plain lacked many stone and metal raw materials that appear in archaeological assemblages. Movement or exchange connected floodplain settlements to foothills, plateau margins and more distant geological sources. River routes were one component of that circulation. The contrast between resource-rich wetlands and relative scarcity of knappable stone in much of the alluvium created reasons for mobility even when food production was locally secure. 5.23 Relocation need not mean social collapse When erosion removes a settlement edge or a channel occupies former fields, households may rebuild nearby while maintaining kin, ritual and economic connections. Archaeologically this can look like abandonment of one mound and foundation of another. Without independent evidence, such movement should not be labelled invasion, depopulation or “collapse.” In a mobile floodplain, relocation can be a normal strategy of continuity. 5.24 Eastern Nepal: the Sapta Kosi fan extends the same problem across the modern border The Sapta Kosi forms a large alluvial fan after leaving the Siwalik Hills. Thakur and Tamrakar described Holocene fan deposits in eastern Nepal ranging from gravel to fine sand and mud and emphasized channel shifting, flooding, erosion and waterlogging as active geomorphic problems (Thakur and Tamrakar 2002). These observations matter to a history of Mithila in Nepal because the modern international boundary does not divide the sedimentary system. Northern and southern parts of the foreland must be analysed together while keeping their political histories distinct. GAJENDRA THAKUR 5.25 Patu and the Rato Khola: foothill terraces offered a different landscape from the deep plain Corvinus’s Patu industry in Mahottari was found on older terrace surfaces near the Rato Khola where erosion exposed abundant lithic material. The foothill/Siwalik setting offered quartzite cobbles and terrace surfaces unlike the fine alluvial tracts farther south. Patu therefore demonstrates both environmental diversity and discovery bias: the same erosion that threatens a surface can reveal a prehistoric workshop. It should not be used as evidence that the contemporary political or cultural identity of Mithila existed in the early Holocene; it is evidence for human use of a landscape that later formed part of the Mithila region (Corvinus 1987, 1989, 2007). 5.26 Anga and the eastern plain: rivers connected rather than isolated the Bhagalpur zone Oriup and Champa belong to a different part of the river system from the north Bihar fan-interfan tract, but they reinforce the importance of river corridors. The Ganga and its tributaries connected fishing grounds, settlements and exchange routes through the Bhagalpur region. Chapter 4 already identified direct fishing technology at Oriup. Chapter 5 adds the spatial point: aquatic resources and movement along the river were part of settlement logic, while channel change could alter access and site preservation. 5.27 Population estimates require more than site counts An increase in the number of known Chalcolithic or early historic sites may reflect genuine demographic growth, but it can also reflect longer occupation, larger settlements, better archaeological visibility or uneven survey. Settlement size, duration, contemporaneity and buried-site probability all matter. This chapter therefore avoids converting raw site totals into population figures. Demographic reconstruction becomes stronger only when survey coverage, chronology and landform history are controlled together. 5.28 Conclusion: settlement was durable because it could move The apparent paradox of the alluvial plain is that long-term human continuity depended partly on flexibility. Rivers supplied water, fish, fertile sediment and routes, but they also eroded, flooded and shifted. Communities managed this uncertainty through location choices, diversified land use, seasonal movement, exchange and, when necessary, relocation. The historical unit is therefore not the immobile village but the settlement landscape: houses, fields, wetlands, paths, grazing areas, channels and remembered places that changed in relation to one another. This conclusion prepares the transition from prehistoric environmental adaptation to the more densely documented social and economic institutions of later chapters. Table 5.1 — Environmental setting and settlement implications Landscape / process Potential Potential costs Historical caution advantages Natural levee / raised better drainage; route bank erosion; crevasse raised ground is not bank access; proximity to flooding; channel flood-free ground water migration Backswamp / oxbow fish, reeds, water, waterlogging; wetland value varies grazing, clay mosquitoes; limited by season and house sites technology Older alluvial surface longer soil greater distance from older surface does not development; relative active water in some automatically mean stability settings older occupation HISTORY OF MITHILA, VAJJI & ANGA — VOLUME II Landscape / process Potential Potential costs Historical caution advantages Young fan / splay fresh sediment; new erosion, sand low site visibility may surface land; changing deposition, unstable reflect burial channels drainage Foothill / terrace edge stone raw material; erosion; local tectonic cannot be treated as drainage; varied and slope hazards identical to low ecotones alluvial plain Avulsion corridor new water and rapid relocation of abandoned channels sediment pathway channel and routes need independent dating Interfan plain mixed river and frequent overbank small rivers can be wetland resources flooding and highly mobile too waterlogging PART II EARLY HISTORIC SOCIETY: VIDEHA, VAJJI AND ANGA Chapters 6–18