Ancient plant remains do not speak for themselves. Archaeobotany teaches learners to connect seeds, chaff, charcoal and microscopic plant evidence with excavation context, preservation, recovery methods and carefully bounded questions about past human activities.
Our original exercise compares two fictional sediment samples: 120 charred grains in twenty litres and 55 grains in five litres. The second has a higher recorded grain concentration even though the first has more grains in total. These counts are invented, as are the archaeological settings. The featured scientific archaeobotany figure from Pangwari by Louis Champion and colleagues is a separate real, CC BY 4.0 source.
The Historic England Archaeobotany team, its 2025 environmental archaeology guidance and eduKate’s Gold Standard hub provide the source and teaching framework.
1. Begin with the recovered seed, not an agricultural legend
A fictional archaeological hearth contains a dark plant fragment. That find does not immediately prove that people grew a named grain, traded it or ate it every day. The material might be a charred seed, a wood fragment or something entirely different. Its excavation context could be disturbed or mixed. Archaeobotany investigates plant remains in archaeological settings and reconstructs human–plant relationships through preservation, identification and context. The first lesson should make students describe the evidence rather than generate a dramatic story from a tiny object.
Use photographs from documented research, not unsupervised excavation materials. Have pupils record context, physical appearance, possible preservation state and what needs specialist examination. Our classroom site, hearth, species and numbers are invented. The genuine featured image is a scientific archaeobotanical figure from Pangwari, credited to Louis Champion and colleagues under CC BY 4.0; it is not a photograph of the fictional hearth.
2. Distinguish macro- and microbotanical evidence
Seeds, grains, chaff, fruits, charcoal and larger wood pieces can provide macroremain evidence. Pollen, phytoliths and starch grains represent different microscopic evidence types. Their sampling, identification and interpretations are not identical. A starch grain or phytolith should not be silently entered into a table of charred cereal grains, and the presence of pollen does not necessarily prove cultivation of its plant.
Show a modern diagram or published micrograph alongside photographs of larger botanical remains. Students identify the material category and the kind of specialist comparison it requires. Historic England’s Archaeobotany team describes these categories and the different ways plant materials can inform food, fuel, construction and past landscapes. Technical terms become useful when they prevent false comparisons.
3. Learn how preservation shapes the archaeological record
Some plant remains survive because they were charred, waterlogged, mineralised or desiccated. Others decay and leave little recoverable evidence. The resulting plant assemblage does not provide a complete inventory of everything once grown or consumed. A settlement without recovered charred peas might still have used peas if they were not exposed to preservation conditions or missed by the sampling programme.
Give students two invented deposits: a burnt hearth layer and a waterlogged ditch. Ask what categories of remains might survive and which would be difficult to compare directly. Waterlogged remains may require different conservation care, while charred material can still be fragmented or overlooked. The lesson is that negative evidence needs a stated expectation of what could reasonably have survived and been detected.
4. Record sediment volume and archaeological context
A sample’s source layer, archaeological feature, stratigraphic position, processing method and sediment volume are part of the botanical evidence. Twenty seeds recovered from twenty litres and twenty recovered from one litre are not the same concentration. Nor does a midden sample necessarily reflect the same activity as a sealed storage vessel. Context and formation processes determine how botanical measurements can inform past behaviour.
Build a fictional sample card including context ID, sampled litres, deposit description, material classes and recovery notes. Ask whether a bag labelled ‘many grains’ would be reproducible. It would not. A proper record lets another researcher check the count, denominator and interpretation. This links Archaeobotany with Teaching Archaeology without repeating a general excavation course.
5. Explain flotation without promising perfect recovery
Flotation separates some lighter organic remains from archaeological sediment by appropriate water-based recovery under specialist oversight. It does not recover every plant type with identical efficiency. Soil properties, mesh selection, fragment density, breakage and the residue-processing workflow influence what is found. A flot should therefore be interpreted alongside the sample history, including any heavier fraction that was not fully examined.
Historic England’s 2025 Environmental Archaeology guidelines explicitly discuss flotation and the need to check residues. Give pupils a diagram of sample, flot and residue, then ask where remains could be missed. Clean modern classroom materials may demonstrate the basic idea when authorised, but genuine archaeological sediments should not be processed outside professional site procedures.
6. Recognise differences among recovery strategies
Hand collection tends to recover larger noticeable finds, while fine sediment sampling and appropriate screening can reveal smaller seeds or fragmented remains. Comparing samples gathered under inconsistent recovery strategies without qualification is misleading. A bag of large fruit stones picked up individually and a flotation sample full of small grains might reflect methods rather than different ancient economies.
Have learners examine fictional records of hand-collected specimens and systematically sampled sediment. They should note what each method could omit and why the two counts are not directly comparable. Ask whether all contexts at the site were sampled in the same way. A consistent workflow may support comparisons; an uneven one needs uncertainty notes or a restricted research question.
7. Identify taxa using comparative collections
Specialists examine plant shape, size, surface characteristics and diagnostic structures and compare them with reference collections. Charring can distort remains, and some fragments cannot be assigned securely to one species. A broad identification such as ‘cereal grain’ can be better than an unjustified exact Latin binomial. A learner should know the difference between ‘observed diagnostic feature’, ‘candidate taxon’ and ‘secure identification’.
Use a licensed image displaying several charred seeds, then ask which traits are visible and what a microscope or comparison specimen could add. Historic England maintains thousands of reference seed and fruit specimens. Do not simply rename the real Pangwari scientific figure’s taxa to fit our imaginary hearth. An auditable report preserves the published identification and records any uncertainty in the classroom analyst’s own exercise.
8. Distinguish crops, weeds, chaff and fuel
An assemblage containing cereal grains, weed seeds and chaff can support questions about processing and agriculture. Yet a wild seed may have been food, fuel contamination, an unintentionally harvested weed or part of a natural background deposit. The label ‘weed’ is relational to a farming system, not evidence the species had no other cultural use.
Supply a fictitious sample containing clean grain, husk fragments and several wild seeds. Ask what evidence would support storage, milling, threshing or dumping. Deposit location, tools and related stratigraphy might help. Students should not infer all plants were locally cultivated from the mixture alone, because trade and secondary disposal remain possible.
9. Use counts per litre instead of misleading raw totals
In our invented exercise, a domestic layer yields 120 identified charred grains from twenty litres of processed sediment, while a refuse feature yields 55 grains from five litres. The first has more grains, but the concentrations are 120 ÷ 20 = 6 per litre and 55 ÷ 5 = 11 per litre. Under consistent recovery conditions, the second has the higher observed concentration.
The calculation is an interpretation of this artificial dataset, not a historical farming result. A higher concentration does not show a greater share of calories, local crop productivity or household preference. Students must say what counted as one grain and which volume entered the denominator. If one context was sampled with a finer screen, the interpretation requires further qualification.
10. Distinguish taxon ubiquity from abundance
A crop can appear in many archaeological samples with few grains each, or in just one sample with an unusually large concentration. Sample ubiquity measures the fraction of examined samples in which a taxon is recorded; it does not directly measure the fraction of inhabitants who ate it. Five comparable samples with barley in four have 4/5 = 80% sample ubiquity.
Ask students to compare barley present in four samples with a legume present in only one, and then compare total counts. The two measures answer different questions. A researcher should disclose how many samples were examined, whether volumes were comparable and whether absences reflect recovery failures. An apparently compelling percentage has little historical meaning without its sampling basis.
11. Interpret wood charcoal in the right context
Wood charcoal may indicate fuels, burnt structures, industrial processes or redistributed debris. The same taxon could have arrived as locally gathered branches, imported timber or reused building wood. Charcoal identification can help with forest history but a small hearth assemblage is not a straightforward survey of the surrounding woodland.
Give learners a fictional hearth with oak charcoal and charred cereals. Ask which evidence supports burning oak wood and which would establish that oak dominated the local forest. Independent pollen, vegetation, archaeological context and appropriate sampling may be needed for the second. The educational result is a sourced hypothesis rather than an unsupported reconstruction of an entire environment.
12. Distinguish food preparation from storage and refuse
Charred grain in a sealed burnt storage feature may tell a different story from scattered seeds in mixed refuse. Burning, movement, trampling, dumping and later intrusion can reshape assemblages. A student who interprets every grain in a hearth as a meal eaten that day misses these depositional processes.
Use two imaginary contexts with the same taxon but different positions: a vessel beneath a collapsed roof and a pit with several layers of mixed refuse. Learners should explain why the storage hypothesis is stronger in one and why a single deposition episode cannot be assumed in the other. This teaches interpretation from botanical and archaeological evidence together.
13. Use radiocarbon dating with source-context caution
Charred short-lived seeds can be useful dating samples when securely related to the archaeological activity of interest. However, a radiocarbon result dates the carbon in sampled material under the method’s assumptions, not automatically a building’s construction or an entire site’s abandonment. Intrusion or redeposition can break the intended event association.
Historic England’s radiocarbon modelling guidance examines suitable sample selection and the old-wood issue. Have pupils state which precise event the seed might relate to and how excavated context supports that association. They should not propose unsupervised destruction of heritage samples to obtain a date.
14. Make contamination and data provenance explicit
Modern seeds can enter deposits through roots, animals or recent soil movement, while laboratory procedures may add further sources of accidental mixing. A database therefore needs site, context, sample size, processing fraction, identification status, analyst and quality notes. A spreadsheet with a neat botanical tally but no context IDs cannot support a defensible history.
Add one obviously modern-looking seed to a fictitious ancient sample and ask students what should happen next. The appropriate response is to flag the observation and seek specialist review, not automatically remove it to make the assemblage look pure. If the identification or context is uncertain, the published table should disclose that uncertainty.
15. Avoid converting remains into a universal diet
Ancient plant remains can inform farming, food choice, fuel use, medicine, craft and trade, but the surviving material reflects preservation, burning, sampling and site histories. One sample cannot establish the percentage of a whole population that ate barley or the exact productivity of a field. Historians need different sources and sometimes models to investigate those broader claims.
Give learners the calculated concentrations and ask them to write the strongest descriptive sentence and a cautious interpretive hypothesis. The correct description states recovered grains per litre; the hypothesis explores refuse concentration or burning history with alternative explanations. The student should learn that measurement quality and interpretive restraint reinforce one another.
16. Assess transfer with different preservation and recovery
The final packet contains a fictional waterlogged ditch and a charred hearth, processed under different methods, with unequal sample volumes and uncertain taxonomic identifications. Students decide which measures are comparable, calculate only justified concentrations and describe how recovery and preservation affect the interpretation.
The strongest answer may refuse a direct numerical comparison while explaining what an archaeobotanist would need to make one. The work should include context numbers, units, methodological notes and evidence of independent judgement. That is the gold standard: a recoverable plant history rather than an invented account of ancient agriculture.
Worked sample comparison: calculate before explaining
All values are fictitious and assume that the same type of identifiable recovered charred grain was counted in the stated processed volume. No numerical result describes a real community.
| Measure | Context A: invented occupation layer | Context B: invented refuse feature |
|---|---|---|
| Processed sediment | 20 L | 5 L |
| Charred grains recovered | 120 | 55 |
| Recovered grains per litre | 120 ÷ 20 = 6 | 55 ÷ 5 = 11 |
| Historical interpretation | Recovery, deposit formation, preservation and species identification require separate review | |
Discussion: A has the larger raw total, while B has the larger observed concentration. Which supports more grain consumption? Neither calculation alone. The class should ask how both deposits formed and whether their sampling and processing methods are sufficiently comparable.
A second case: 80% ubiquity is not an 80% diet
Four of five invented samples contain barley. Sample ubiquity is 4 ÷ 5 = 80%; that does not show that 80% of people ate barley. It reports only the proportion of examined samples with recorded barley under the sampling and identification rules. Ask learners how selective sampling or different grain preservation would change the result.
Four lessons with a new transfer task
- Lesson 1: describe plant-remain types, preservation and archaeological context.
- Lesson 2: identify sampling and recovery limitations; compare diagnostic botanical features.
- Lesson 3: calculate concentrations and ubiquity with complete numerator and denominator definitions.
- Lesson 4: interpret an unfamiliar assemblage with contrasting preservation and methodology.
A preservation audit: why the same crop can appear differently in two deposits
Imagine one fictional store destroyed by fire and another storage area gradually abandoned without burning. The first may preserve charred grains, while the second may not preserve the same plant parts at all. An archaeologist recovering many charred grains from the fire cannot conclude that the inhabitants there ate more of that crop than those using the unburnt store. The survival conditions are radically different. Even within the burnt deposit, the distribution of seeds may reflect collapse, sweeping, sampling or subsequent disturbance rather than the stored quantity.
Ask pupils to draw two separate chains from human activity to deposit, preservation and modern recovery. The chain for the burnt store may contain a burning event that transformed organic material, while the other chain may end in decay or waterlogging. For each, identify what is directly documented and what remains a hypothesis. A strong answer explains that the archaeological assemblage is a selective record, not a complete historical pantry. The educator can introduce a new context and check whether the learner revises the preservation model instead of simply repeating the original crop story.
Recovery audit: a hidden selection bias in the flotation fractions
The invented context A was sampled at twenty litres, but suppose only its floating fraction was carefully sorted. Context B used a different procedure that also examined larger portions of the heavier residue. Their recorded concentrations—six and eleven recovered grains per litre—are still mathematically correct under the exercise’s counted material. However, the original deposits may have had different recovery efficiencies. The smaller number from A could reflect missed material as well as a genuine difference in the ancient activities being sampled. No safe fixed multiplier can be guessed from the two totals alone.
Students prepare a method card recording collected volume, processed volume, screen or fraction information, preservation state and which residues were examined. They should propose what specialists might check to assess recovery rather than improvise archaeological laboratory steps without supervision. The Historic England environmental-archaeology guidance provides a professional reference on recovery limitations. The teaching standard is that a student can distinguish arithmetic errors from methodological incomparability; correcting the calculation alone will not remove sampling bias.
Taxonomic uncertainty: why a broader identification may be a better result
A tiny carbonised fragment looks superficially like a wheat grain, but the provided image does not preserve the features needed to distinguish it confidently from other cereals. A student may select wheat because the fictional story mentions bread. The better answer is to state the identifiable features and give the broadest category the evidence securely supports. Professional collections and reference specimens can help, but a low-quality source image should not be made to carry an unsupported species assignment.
Change the hypothetical dataset so ten fragments previously labelled wheat are reclassified as indeterminate cereals. The total number of plant objects may remain the same, while wheat-specific abundance changes. Ask which historical conclusions must be revised. If the original argument was that the community cultivated wheat locally, the revised identifications weaken one part of that chain; they do not prove a different crop was cultivated. Students should preserve identification history and note who revised it. This is an evidence-management lesson as important as looking at botanical shapes.
Two measures: abundance and ubiquity are not interchangeable
Create five imaginary bulk samples from one site. Barley appears in four, with a total of only twenty identifiable grains. Legumes appear in one sample, with 200 grains concentrated in a particular deposit. Barley ubiquity is four-fifths, or 80%, and legume ubiquity is one-fifth, or 20%. The raw legume count is nonetheless ten times larger. Which plant was most important in daily food? Neither measure can answer that question alone without contextual and behavioural evidence.
Ask students to explain the two denominators out loud: counted botanical remains for abundance, and examined sample count for ubiquity. Then reveal that the one very rich legume sample came from a burnt store while all the barley samples came from mixed refuse. The formation histories differ further. Pupils should rewrite their captions to report exactly which statistical claim is supported. This exercise demonstrates why an impressive number can be technically correct but historically misleading if the source population and depositional context are not stated.
From crop-processing remains to a proportionate historical explanation
A fictional occupation area contains a mixture of grain, chaff and small wild plant seeds. One student announces that threshing took place there; another says the seeds are proof of imported grain. Both are possible questions, but neither is established by the list alone. Chaff may be related to crop cleaning, yet plant processing, storage, dumping, fuel and later movement can overlap. Archaeobotanical reasoning depends on identifying the materials correctly and tracing how they entered the deposit.
Give teams three evidence cards: an associated tool, the spatial position of the sample, and a comparison with nearby features. Ask which source would strengthen a hypothesis of local processing and which would remain ambiguous. A pupil should be able to state that the assemblage is consistent with plant-processing debris without claiming every grain was grown on an adjacent field. This nuanced explanation still teaches genuine human history; it simply refuses to invent a complete agricultural system from a small fraction of what survived.
Radiocarbon case: the seed may not date the event you want
A short-lived charred seed can be valuable in radiocarbon research, but its relationship to a particular human event must be demonstrated. A grain from a sealed burning layer may have a different association from a grain lodged in a root channel that cuts through the layer. The radiocarbon method concerns the sampled carbon and chronological modelling; it does not automatically date the entire building, activity phase or community. A defensible source claim requires secure provenance and specialist dating interpretation.
Ask learners to choose between two invented sampling candidates and explain their reasoning without proposing to damage real heritage material. They should say what excavation record could establish whether the seed belonged to the original deposit and how later movement might affect interpretation. If the association is unresolved, even a technically precise radiocarbon result may be inappropriate for dating the event of interest. This distinguishes the measurement’s analytical quality from the historical validity of the selected sample.
Final assessment: a new assemblage with unequal preservation and sampling
For the transfer task, replace the initial pair of charred contexts with a waterlogged ditch, a burnt storage layer and an exposed surface disturbed by later roots. Their plant fragments differ in preservation, identification certainty, sample volume and collection method. Learners prepare a table recording which comparisons remain valid, which must be restricted, and which potential historical behaviours could account for the pattern. The teacher should not supply the preferred farming narrative in advance.
Assess four progress stages: recognition of plant types and preservation; accurate sample and count recording; appropriate calculation of concentrations and ubiquity; and independent interpretation that acknowledges alternative deposition or recovery processes. Students who decline an invalid cross-context comparison may have demonstrated more advanced reasoning than students who calculate a percentage without checking its denominator. The gold standard is a reproducible chain from sample context to defensible, revisable historical explanation.
Frequently asked questions
Does one charred seed prove a local crop was cultivated?
No. Context, identification, deposition and possible import are separate research questions.
Does flotation recover everything?
No. Recovery and processing can selectively miss remains; residues and methods matter.
Can higher grain counts establish a richer diet?
No. Sample size, preservation and deposit function must be considered.
May students handle archaeological samples without authorisation?
No. Use published material, clean simulations and supervised professional protocols.
Professional and internal references
- Historic England: Archaeobotany
- Environmental Archaeology: Third Edition
- Historic England: Environmental Archaeology
- Pangwari Archaeobotanical Figure (CC BY 4.0)
- Teaching Agricultural Science
- Teaching Archaeology
- Teaching Geoarchaeology
- Gold Standard master hub
The gold standard is a student who preserves the context in which a plant fragment was found, explains how it was recovered and identified, and makes no historical claim that extends beyond the samples and evidence.