THE INDUS CIVILIZATION | ARCHAEOLOGICAL RECONSTRUCTION

THE INDUS CIVILIZATION

Settlements, Hydrology, Production Networks and Population Movement

A source-critical archaeological reconstruction

AUTHOR & RESEARCHER SHUBHAM PHOUGAT

Evidence boundary: settlement and material culture can be reconstructed in considerable detail; the Indus sign system remains undeciphered, and its political, linguistic and religious meanings cannot be read as if they were known.

Research report by Shubham Phougat • 1

Executive summary

What the archaeological record allows us to reconstruct—and what it does not

The Indus or Harappan Civilization was a broad, internally diverse Bronze Age world of towns, villages, craft centres, agricultural zones and cities connected by shared material practices. Its mature urban phase is conventionally dated to about 2600–1900 BCE, preceded by regional and Early Harappan developments and followed by varied Late Harappan traditions. The chronology is not a single clock: local sequences and definitions differ across the region. [14, 24–27]

The civilization cannot be reduced to the Indus River valley alone. Settlement evidence extends from Baluchistan and the Makran coast through the Indus basin, Punjab and the Ghaggar-Hakra region to Gujarat, Kutch and parts of the upper Ganga-Yamuna region. The broad range reflects ecological adaptation and dense regional connections, not necessarily one unified state controlling every settlement. [1, 3, 14, 24–26]

The report’s central finding is that Harappan urbanism depended on multiple connected systems: floodplain and rain-fed agriculture; varied crops and livestock; local water harvesting; craft specialisation; standardised measures; exchange in stone, metals, shell and finished goods; and movement between towns and hinterlands. No single river, crop, trade route or political centre explains the whole system. [1–4, 9, 12–16]

Three limits govern the reconstruction. First, site discovery and excavation are uneven, so maps are records of archaeological knowledge as much as of ancient settlement. Second, hydrological reconstructions are robust in broad outline but some channel histories and the timing of climatic stress remain debated. Third, the Indus sign system is undeciphered: statistical regularities can support competing interpretations but cannot by themselves reveal a language, a king list, a theology or a political constitution. [1–3, 5–7, 17–22]

1. Method and evidence hierarchy

This report treats the Indus as an archaeological problem, not as an exercise in identifying modern national, linguistic or religious identities with Bronze Age populations. The main evidence classes are: (1) excavated settlement plans and stratigraphy; (2) radiocarbon and optically stimulated luminescence (OSL) chronologies; (3) geomorphology, palaeochannel sediment provenance and palaeoclimate proxies; (4) archaeobotanical and zooarchaeological remains; (5) raw-material sourcing, craft debris, tools and finished objects; (6) isotopes and ancient DNA; and (7) the sign corpus, whose meaning remains unresolved. Each class answers different questions and has different biases. [1–12, 17–22]

“Evidence hierarchy” does not mean that one type of evidence always outranks another. A palaeochannel core can establish the age and source of sediments, but not directly tell us what a farmer believed. A bead’s geological source may establish exchange links, but not the exact route, trader, price or political treaty. A DNA profile can model ancestry, but does not identify language, ethnicity or cultural identity on its own.

The report uses the following labels: established means directly supported by multiple independent lines of evidence; strong inference means the best explanation of converging data but not directly observable; contested means credible specialists disagree or the evidence is not unique; speculative means a possible interpretation that lacks decisive independent support.

2. Chronology: a regional process, not a sudden urban invention

Figure 1. Conventional broad phases; local chronologies vary and the antecedent village traditions are not identical to the later urban

civilization.

Regional antecedents include early farming and herding communities such as Mehrgarh in the Kachi Plain, occupied from the seventh millennium BCE. Over millennia, farming, craft production, settlement permanence, exchange and regional styles developed across northwest South Asia. Early Harappan communities (often dated c. 3300–2600 BCE) show increasingly connected settlement and material patterns; the Mature Harappan phase (c. 2600–1900 BCE) exhibits the strongest interregional standardisation and large urban centres; after c. 1900 BCE, many cities declined or changed, and regional traditions diversified. [14, 24–27]

This sequence avoids two errors. The civilization did not appear fully formed at a single moment; nor did the end of the mature urban system mean that Harappan-descended people vanished. Post-urban sites often show continuity in pottery traditions, subsistence and settlement, while other places were abandoned or saw reorganisation. “Collapse” is sometimes used as shorthand, but “deurbanisation and regionalisation” better describes the uneven archaeological pattern. [1, 3, 5, 10, 14]

3. Mapping the settlement system

Geographical distribution, density, settlement hierarchy and archaeological visibility

Figure 2. Schematic locator map of selected major and regional sites. Site positions are approximate; rivers and palaeochannels are

generalized, and the map is not a surveyed GIS layer.

The settlement map is best read as a network of environmental zones rather than a neat river corridor. Harappa sat near the Ravi; Mohenjo-daro on the lower Indus; Rakhigarhi and other sites occupied the Haryana/upper Ghaggar-Hakra region; Ganweriwala and many settlements lay in Cholistan; Kalibangan lies in Rajasthan near the palaeochannel system; Dholavira occupies Khadir Island in the Rann of Kachchh; Lothal sits in coastal Gujarat; Mehrgarh represents much earlier farming and craft traditions in Baluchistan; and Shortugai in northern Afghanistan represents an important northern outpost often discussed in relation to resource and exchange networks. [3, 13–16, 24–26]

Site-size categories are not straightforward. “City”, “town”, “village”, “craft centre” and “outpost” are analytical labels that depend on excavated extent, density, functions, and dated occupation. A large mound is not automatically fully urban throughout its history; a smaller site can be economically important if it produces a specialised commodity or occupies a strategic route. Known-site distributions also reflect modern survey intensity, land-use change, erosion, alluviation and political access. [3, 11, 13, 24, 25]

4. Hydrology and palaeolandscapes

Why the Indus network cannot be mapped by simply drawing modern rivers

The Indus system is a high-energy river network shaped by Himalayan snow and ice melt, monsoon rainfall, tributaries, sediment loads, avulsion and floodplain aggradation. In Punjab, the Indus and its tributaries create fertile incised valleys but also unpredictable flood risks. In lower Sindh, the distributive alluvial system raises and shifts channels. In arid and semi-arid regions, seasonal streams and groundwater can matter as much as large rivers. Settlement placement therefore reflects a moving relation between water, cultivable land, flood risk, transport and local storage—not proximity to a single static river. [1, 2]

The Ghaggar-Hakra palaeochannel debate is especially consequential. Geological and sediment-provenance studies identify parts of the buried channel system as an older course of the Sutlej, but dating indicates that a major Himalayan-fed flow had shifted away before mature Harappan urbanism. Fine sediments and channel traces indicate that seasonal and monsoon-fed flows continued in portions of the old valley after the large river moved. This allows a densely settled Holocene landscape without requiring a large perennial glacier-fed river to have flowed through every Harappan-period settlement zone. [1, 2, 11]

This does not mean that every question is closed. The system is spatially complex; different reaches can have different histories, and geomorphological interpretations depend on the location and dating of cores, the provenance methods used and the age model. Some studies address palaeo-Yamuna connectivity and older phases, while others focus on later Holocene sediments. A claim about a channel during the Pleistocene cannot automatically be transferred to 2600 BCE. [1, 2, 11]

Hydroclimatic stress likely increased Monsoon proxies, lake records, river sediment Did climate change cause urban decline? vulnerabilities, but urban change was

Identifying a textual river name with a specific Textual traditions, historical geography and ancient channel is a multi-step historical What did “Sarasvati” refer to? palaeochannel studies. hypothesis, not something geology can determine alone.

5. Agriculture and food security

A diversified system across rainfall and flood regimes

Archaeobotanical remains indicate varied crops across Indus regions and periods: wheat and barley; pulses such as lentil, pea and chickpea; oilseeds; sesame; cotton; and, in different settings, millets and rice. The old simplifying model in which the “core” grew winter cereals and the “periphery” grew summer crops is too neat. Crop assemblages show multi-cropping and adaptation to local environments; water availability and seasonality constrained what could be grown where. [4, 16, 27, 28]

Crop evidence is also an archaeological sample, not an exhaustive inventory. Plant remains survive unevenly. Only a portion of known settlements have archaeobotanical samples, and systematic flotation and full publication have historically been uneven. Presence confirms use; absence can be preservation or sampling bias rather than proof that a crop was unknown. [4, 16, 17]

A plausible agrarian system combined winter and summer cropping where conditions allowed, flood-recession farming, local rainfall, livestock, foraging and exchange. Water management at Dholavira is a notable case: reservoirs, channels and other installations are adapted to a relatively water-stressed setting. Its design demonstrates local ingenuity, not a universal template for every Harappan city. [14–16]

After c. 1900 BCE, several regional sequences show changes in crop mix and growing settlement numbers in some eastern areas. Greater use of drought-tolerant crops and altered crop calendars could be one means by which households managed climatic and economic uncertainty. Such choices should be read as adaptation and reorganisation as well as stress. [4–7, 10]

6. Craft production and technological specialisation

The material record shows sophisticated production in ceramics, stone tools, beads, shell ornaments, faience, copper/bronze objects, seals and weights. Specialisation did not require a single central bureaucracy to control every workshop. At Harappa, pottery production research identifies skilled sequences, repetition and standardisation, while finding limited direct evidence for a single central authority controlling production. Different crafts may have been organised through households, kinship groups, neighbourhoods, attached specialists, itinerant artisans or larger exchange relationships. [12, 24, 25]

Chanhudaro is an important craft-centre case because excavated evidence includes bead-making, shell-cutting, metal-working and seal/weight production. At Dholavira, production and raw-material handling reveal ties to regional supply systems. Carnelian and agate production links are commonly associated with Gujarat/Kutch; shell-working is found at coastal sites; chert from the Rohri Hills was an important toolstone; copper arrived from multiple possible source areas including Rajasthan/Aravalli and farther exchange systems; and lapis lazuli from the Badakhshan region connected the Indus world with distant highland networks. Specific source assignments vary by material and study; all routes should not be assumed equally certain. [12, 13, 23–26]

Standardised weights, brick proportions in many contexts, seal forms and craft techniques provide evidence of shared conventions across a wide area. The crucial inference is shared measurement and practice; the stronger claim that these necessarily indicate one imperial state is not warranted by standardisation alone. Similar standards can emerge through exchange, guild-like institutions, political arrangements, shared metrology or overlapping regional authorities. The unreadable sign corpus prevents a direct account of the institutions that administered these systems. [12, 24, 25]

Figure 3. A conceptual production-and-exchange network; arrows are schematic, routes are inferred at different confidence levels, and the

diagram does not imply a single central authority.

7. Exchange beyond the settlement network

Regional circulation, Gulf routes and the Mesopotamian connection

Long-distance connections with the Persian Gulf are supported by more than one evidence type. Indus-style seals and other objects occur in Mesopotamian contexts; Mesopotamian texts refer to a land called Meluhha, widely associated by scholars with the Indus region, alongside Dilmun and Magan in Gulf exchange networks. Archaeological evidence includes sealings, beads and materials whose style or provenance connects communities across distance. Such objects show interaction, but not necessarily direct voyages by every producer or a single trading monopoly. [23, 24, 28]

An exchange chain may have involved several kinds of intermediaries: local producers, inland carriers, coastal communities, merchants at Gulf hubs, and merchants in Mesopotamian cities. The exact relationship between object, merchant and route is often not recoverable. Isotopic sourcing of artefacts and human remains can help distinguish origins and mobility, while ship technologies, harbours and coastal geography inform route hypotheses. The proposed “dockyard” at Lothal is famous but its precise function remains disputed; its basin and water-related architecture should not be treated as a final answer to every maritime question. [13, 23, 28]

The Indus trade system should not be understood as a modern market economy or a simple state monopoly. We have limited evidence for prices, contracts and merchant institutions compared with Mesopotamia, where cuneiform documents often disclose individual transactions in detail. Indus weights, seals and craft-standardisation point to exchange protocols, but their administrative meaning remains uncertain. [12, 18–20, 23–25]

8. Population movement and biological evidence

What isotopes and ancient DNA can show—and what they cannot

Population mobility can be investigated through material evidence, cemetery patterns, strontium and lead isotopes in tooth enamel, oxygen isotopes, and ancient DNA. Enamel forms in childhood and may preserve a chemical signature associated with the geology and water of a person’s early residence. It does not provide an exact birthplace unless local baseline values and sampling are sufficient. PLOS ONE research on Harappa and Farmana identified isotopic patterns consistent with immigration from resource-rich hinterlands for some cemetery individuals. The study also stresses that formal cemeteries contain only a small and potentially selective sample of the population. [6, 11]

The 2019 Rakhigarhi genome study attempted to recover DNA from 61 skeletal samples and obtained one sample of sufficient authenticity and quality. The individual had ancestry related to a population cline found among certain people at sites in cultural contact with the Indus world, and lacked evidence of Steppe pastoralist-related ancestry. This was an important result, but one individual is not a representative census of an enormous and diverse civilization. The same paper explicitly calls for more ancient DNA across the Indus range. [8]

Narasimhan and colleagues’ wider study of 523 ancient individuals across South and Central Asia provides a broader model of ancestry and movement, including later spread of Steppe-related ancestry into South Asia in the second millennium BCE. The dataset is not 523 Harappan genomes; most are from other regions and periods. Genetic ancestry is not the same as language, culture, caste, religion or political identity. Genetic results constrain demographic models but do not by themselves establish what language a Harappan spoke. [9]

9. The 1900 BCE transition: deurbanisation, not disappearance

Around the end of the Mature Harappan phase, many large urban centres contracted, some settlements were abandoned, standardised artefact traditions changed and long-distance trade appears to have been reorganised. This did not occur

simultaneously or identically everywhere. Some communities continued at old locations; others shifted toward more eastern or southern regions; small settlements and regional traditions became more prominent. [1, 3–7, 10]

Hydroclimatic change is a major part of the explanation. Giosan et al. argue that declining monsoon strength and fluvial quiescence made some floodwater farming systems less dependable and increased vulnerability. Other palaeoclimate research explores one or more episodes of reduced winter and summer rainfall. But recent reviews stress disagreements over the timing, regional coherence and causal strength of a single “4.2 ka event.” Environmental stress is best treated as a changing set of constraints that interacted with agricultural strategy, population movement, exchange, local institutions and urban organisation. [1–7]

The map of Late Harappan sites in northwest India suggests density may have increased in particular zones even as large cities declined. Green and Petrie’s settlement-database approach cautions that results depend on site discovery, period assignment and recording practices, but it illustrates why “collapse” can mask a shift in settlement size and distribution rather than simple depopulation. [3]

A balanced causal account therefore includes several interacting processes: weaker or less reliable monsoon conditions; changes in flood intensity and channel behaviour; crop diversification; limits of large-city provisioning; shifts in regional exchange; changes in craft and administrative coordination; and migration or resettlement. It remains difficult to quantify the weight of each factor for a specific city, especially where excavation histories and environmental records are incomplete. [1–7, 10]

10. The Indus sign system: explicit limits of interpretation

The sign system often called the Indus script remains undeciphered. Thousands of inscribed objects are known, including seals, sealings, tablets, pottery and other small objects, but most inscriptions are very short compared with the long texts available for deciphering Egyptian hieroglyphs or cuneiform. No generally accepted bilingual inscription links Indus signs to a known language and an independently readable text. The corpus is therefore too limited to yield secure translations of royal titles, laws, myths, personal names or theological doctrines. [18–22]

The central debate includes: (a) a linguistic hypothesis, in which the signs encode a language; (b) interpretations that treat the system as a nonlinguistic or only partly linguistic sign system; and (c) intermediate models that allow signs to record names, status, identity, commodities or ritual information without a full text comparable to later writing. Rao et al. (2009) reported statistical properties consistent with linguistic structure. Sproat (2014) argued that the statistical methods then used did not discriminate language from nonlinguistic systems reliably and classified the tested Indus corpus as nonlinguistic by his measures. Farmer, Sproat and Witzel (2004) advanced a broader critique of the literacy hypothesis. These positions remain debated; none should be represented as a universally accepted final decipherment. [18–20]

Current computational work does not change this core limit. Statistical models can measure sign frequency, ordering, repetition and predictability. They cannot supply a unique translation unless hypotheses are independently anchored. Recent 2026 preprints illustrate continuing disagreement over whether statistical features can distinguish a linguistic system from structured nonlinguistic systems. These are preprints, not settled peer-reviewed decipherments, and their contributions should be assessed at the level of method and reproducibility. [21, 22]

Consequently, this report does not attribute a named language, modern ethnic identity, pantheon, priesthood, emperor, caste system or constitutional model to the Harappans on the basis of the signs. Such hypotheses may be explored in scholarship, but they must remain hypotheses until additional evidence—such as a longer inscription, a bilingual text, or a securely anchored sign-to-language relationship—can test them.

11. Institutions, standardisation and political organisation

The Indus Civilization exhibits strong shared conventions across a very large region: weights, brick proportions in many contexts, seal practices, urban planning features and craft procedures. Yet the political system is less visible than in Egypt or Mesopotamia, where royal inscriptions, palaces, king lists, temples and extensive administrative texts sometimes provide direct clues to rulers and state institutions. The Indus record has no accepted readable king list and no secure decipherment of political texts. [12, 24–26]

Researchers have proposed different models: a centralised state, multiple interacting regional polities, corporate institutions, merchant or craft networks, or combinations that changed over time. Shared standards could have been enforced by a state, sustained by trade and social practice, or maintained through several overlapping mechanisms. The lack of conspicuous royal monuments and richly differentiated royal tombs has encouraged alternatives to the classic divine-king model, but absence of a familiar type of monument is not proof that hierarchy or political authority did not exist.

A disciplined conclusion is that coordination was real, but its exact structure remains uncertain. City planning and standardized goods imply collective arrangements and skilled organisation; they do not tell us whether these arrangements were administered by a single sovereign, councils, corporate groups, ritual specialists, merchants or varying combinations. The undeciphered signs are a major evidentiary gap. [12, 18–20, 24–26]

12. Major sites as distinct case studies

13. What is established, inferred and still open?

14. Research priorities and what new evidence would change the picture

The strongest next advances are likely to come from expanding datasets rather than producing a single grand theory.

• Settlement GIS: publish interoperable georeferenced site inventories with phase-specific date ranges, confidence flags, excavated area and survey intensity. This will help distinguish actual settlement-density change from changes in archaeological reporting.
• Hydrology and sediment provenance: obtain more dated cores across individual reaches of the Ghaggar-Hakra/Nara system and evaluate river-source claims reach by reach, rather than using one palaeochannel label for a vast region.
• Ancient DNA: sample more individuals from multiple Harappan sites, regions and burial contexts. The current Rakhigarhi result is foundational, but population variation needs a much broader sample.
• Isotope baselines: expand geological, oxygen and strontium baselines and compare cemeteries, settlements and resource zones; test mobility hypotheses against multiple methods.
• Production networks: integrate petrography, trace elements, sourcing and workshop residues for stone, copper, shell, faience and ceramics, while separating raw-material movement from finished-goods movement.
• Script: create a critically edited open corpus with consistent sign IDs, object provenance, context, date, directionality and uncertainty. Any decipherment claim should publish a reproducible method, an independently testable reading and predictions for unseen texts.
• Social reconstruction: use housing, diet, burial treatment, craft distribution and access to water or storage to test inequality and hierarchy. Avoid reading modern caste, religious or national categories into archaeological evidence without independent support. A longer inscription or a bilingual object could dramatically change script research. More human genomes and more well- contextualised settlements could substantially refine mobility models. Until then, responsible scholarship should keep several models open and state which new discoveries would falsify or strengthen them.

15. Conclusion: a complex society, incompletely legible

The Indus Civilization is best reconstructed as a broad and diverse socio-economic network whose urban phase depended on cooperation across agriculture, water access, craft production, metrology, mobility and exchange. Settlements were distributed across different hydrological and ecological settings; Harappans responded to those environments in more than one way. Their political authority may have operated through institutions different from those most visible in royal Egypt or Mesopotamia, but the exact institutional design remains uncertain.

The river evidence cautions against a simplistic “lost great river caused collapse” story. The Ghaggar-Hakra region contained active channels at different times; a major Sutlej course likely preceded mature Harappan urbanism, while monsoon-fed flow and sedimentation continued in portions of the palaeochannel. Later hydroclimatic changes affected farming and settlement differently from place to place. The end of mature urbanism is better described as a long transition, with deurbanisation and regionalisation, than as an overnight disappearance.

The strongest account of population movement combines archaeology, isotope data and ancient DNA, while clearly stating sample limits. It is plausible that mobility and mixing were important to urban society and to later South Asian population history. It is not legitimate to infer language, caste or cultural identity directly from a genetic component, an artefact style or a modern political narrative.

Finally, the undeciphered sign system imposes a decisive boundary on interpretation. Material archaeology can reconstruct settlement, craft, water management and exchange in considerable detail; it cannot read a royal chronicle, theology or law code that has not been deciphered. The right conclusion is neither that everything is unknowable nor that a preferred modern identity has been proven. It is that many aspects of Harappan life can be reconstructed with care, and the remaining gaps must stay visible until evidence changes them.

Selected bibliography and source links

The list prioritises research literature, primary research repositories, UNESCO documentation and transparent statements of disputed interpretations.

[1] Giosan, L. et al. (2012). “Fluvial landscapes of the Harappan civilization.” PNAS 109(26): E1688–E1694. [open source] [2] Singh, A. et al. (2017). “Counter-intuitive influence of Himalayan river morphodynamics on Indus Civilisation urban settlements.” Nature Communications 8:1617. [open source] [3] Green, A. S. & Petrie, C. A. (2018). “Landscapes of Urbanization and De-Urbanization: A Large-Scale Approach to Investigating the Indus Civilization’s Settlement Distributions in Northwest India.” [open source] [4] Petrie, C. A. et al. (2020). “‘Multi-cropping’, Intercropping and Adaptation to Variable Environments in Indus South Asia.” Journal of World Prehistory. [open source] [5] Jahan, T. & Quamar, M. F. (2024). “The ‘4.2 ka drought event’ and the fall of the Harappan Civilization: A critical review.” Review of Palaeobotany and Palynology 331, 105187. [open source] [6] Staubwasser, M. et al. (2003). “Climate change at the 4.2 ka BP termination of the Indus valley civilization and Holocene south Asian monsoon variability.” Geophysical Research Letters. [open source] [7] Dixit, Y. et al. (2022). “Tropical Indian Ocean basin hydroclimate at the Mid- to Late-Holocene transition and the double drying hypothesis.” Quaternary Science Reviews. [open source] [8] Shinde, V. et al. (2019). “An Ancient Harappan Genome Lacks Ancestry from Steppe Pastoralists or Iranian Farmers.” Cell 179(3):729–735.e10. [open source] [9] Narasimhan, V. M. et al. (2019). “The Formation of Human Populations in South and Central Asia.” Science 365(6457): eaat7487. [open source] [10] “Intensified summer monsoon and the urbanization of Indus Civilization in northwest India.” (2018). Open-access research article. [open source] [11] “Evaluating the connectivity of the Yamuna and the Sarasvati during the Holocene: Evidence from geochemical provenance of sediment in the Markanda River valley, India.” Geomorphology (2022). [open source] [12] PLOS ONE (2015). “Evidence for Patterns of Selective Urban Migration in the Greater Indus Valley (2600–1900 BC): A Lead and Strontium Isotope Mortuary Analysis.” [open source] [13] Wright, R. P. “Patterns of Technology and the Organization of Production at Harappa.” Harappa excavation research archive. [open source] [14] OpenStax (2022). World History, Volume 1 to 1500, “The Indus Valley Civilization.” [open source] [15] UNESCO World Heritage Centre (2021). “Dholavira: a Harappan City,” site nomination and technical documents. [open source] [16] UNESCO World Heritage Centre. “Dholavira: a Harappan City.” World Heritage List entry. [open source] [17] Petrie, C. A. et al. (2025). “Early to Mid-Holocene land use transitions in South Asia: A new archaeological synthesis of potential human impacts.” PLOS ONE. [open source] [18] Rao, R. P. N. et al. (2009). “Entropic evidence for linguistic structure in the Indus script.” Science 324(5931):1165. [open source] [19] Sproat, R. (2014). “A Statistical Comparison of Written Language and Nonlinguistic Symbol Systems.” Language 90(2):457–481. [open source] [20] Farmer, S., Sproat, R. & Witzel, M. (2004). “The Collapse of the Indus-Script Thesis: The Myth of a Literate Harappan Civilization.” Electronic Journal of Vedic Studies 11(2). A major minority position; not consensus. [open source] [21] Nair, A. (2026 preprint). “How Non-Linguistic Is the Indus Sign System? A Synthetic-Baseline Scorecard.” Preprint; not peer reviewed. [open source] [22] Raghavendra, N. (2026 preprint). “On the Non-Specificity of Statistical Measures Used in Script Decipherment.” Preprint; not peer reviewed. [open source] [23] Ratnagar, S. (2004). Trading Encounters: From the Euphrates to the Indus in the Bronze Age. Oxford University Press / Oxford India. [open source] [24] Kenoyer, J. M. (1998). Ancient Cities of the Indus Valley Civilization. Oxford University Press / American Institute of Pakistan Studies. [open source] [25] Wright, R. P. (2010). The Ancient Indus: Urbanism, Economy, and Society. Cambridge University Press. [open source] [26] Possehl, G. L. (2002). The Indus Civilization: A Contemporary Perspective. AltaMira Press. [open source] [27] Fuller, D. Q. & Madella, M. (2002/2006). Studies on Indus agriculture, crop regimes, and archaeobotanical evidence; see also the synthesis in Petrie et al. [open source] [28] Petrie, C. A. & Bates, J. (2017). “Multi-cropping, Intercropping and Adaptation to Variable Environments in Indus South Asia.” Journal of World Prehistory. [open source] [29] “A new approach to tracking connections between the Indus Valley and Mesopotamia: initial results of strontium isotope analyses from Harappa and Ur.” Journal of Archaeological Science. [open source] [30] NCERT. “Bricks, Beads and Bones: The Harappan Civilisation” (Class XII Themes in Indian History; useful teaching synthesis, read with research literature). [open source] [31] India Science Heritage / Government portal. Harappan standard weights, material culture and technology (general reference; interpret claims with archaeological literature). [open source]

Appendix A. Short glossary