Full chapter text
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