Experiment 19: Food Web in a Quadrat — Schoolyard Ecology Survey

How many clover plants live in your school lawn? You cannot count them all — but you do not need to. Ecologists estimate whole populations from small samples using a quadrat: a square frame of known area, placed at random, inside which everything is counted. Scale the sample up by area and you have an estimate for the whole habitat — plus, from the species you find, the raw material to build a food web. The Virtual Lab: Food Web in a Quadrat — Schoolyard Ecology on this page provides a simulated schoolyard; the bench version needs string, stakes and your actual grounds.

The science behind the practical

Sampling works when the sample is representative, and the enemy of representativeness is human choice: we unconsciously throw quadrats where things look interesting. The defence is random placement — throw the frame over your shoulder, or use random coordinates along a tape. Repeat throws (five per area here) average out the patchiness of real habitats. Two numbers matter per species: abundance (mean count per quadrat) and, for plants, percentage cover. Population estimate = mean per quadrat × (total area ÷ quadrat area). Our quadrat is 0.5 m × 0.5 m = 0.25 m², so counts are multiplied by 4 to get density per m².

Comparing microhabitats — open grass, under a tree, near a wall — tests how conditions (light, moisture, trampling, leaf litter) filter which species thrive where. Feeding links between the species found then assemble into a food web: producers at the base, herbivores above them, predators at the top, arrows pointing from food to feeder (the direction energy flows).

Variables, controls and fair testing

Wear gloves, do not touch unknown fungi or insects, and wash hands afterwards. Leave every organism where you found it.

Running the investigation well

Throw, don't place. Record everything inside the frame — including the embarrassing zeros for species you expected; absences are data. Estimate percentage cover of grass, bare soil and leaf litter per throw before counting animals, because your arrival disturbs them — count the mobile animals (ants, beetles) first, in fact, before they scatter. Average each area's five throws before comparing areas. In the simulator's schoolyard, survey a fourth microhabitat of your choosing and predict its richness *before* you sample it.

Worked example: a sample investigation

Mean counts per quadrat (average of 5 throws), quadrat area 0.25 m²:

The open lawn's total area is 240 m².

Analysing the sample data

Step 1 — scale up to the whole lawn. Clover density in open grass: 14.2 per 0.25 m² = 14.2 × 4 = 56.8 plants per m². Whole lawn: 56.8 × 240 = 13,632 clover plants (report as ≈ 13,600 — the estimate does not deserve five significant figures). Ants in open grass: 6.4 × 4 = 25.6 per m² — though ants move, so treat that as "ants active per m² at survey time", not a fixed population.

Step 2 — compare microhabitats. Clover crashes from 14.2 (open) to 2.1 (under tree): shade and tree roots outcompete it. The pattern reverses for woodlice (0.8 open vs 4.6 under tree) and snails — damp, litter-rich shade suits them; litter cover (45% under tree vs 5% open) is the environmental clue. Richness peaks in open grass (7 species) and dips under the tree (5).

Step 3 — build the food web. Producers: clover, daisy, grass. Herbivores: snails eat clover and daisy leaves; ants harvest seeds and also farm aphids; woodlice are detritivores, eating the leaf litter itself — an arrow from litter (dead matter) to woodlice, a link students often miss. A bird or spider eating ants and woodlice would sit at the third level. Every arrow points from the eaten to the eater: energy's direction of travel.

Conclusion and evaluation

Conclusion: microhabitat strongly filters the community: the open lawn held an estimated ~13,600 clover plants and the highest richness, while the shaded, litter-covered ground under the tree favoured detritivores like woodlice (4.6 per quadrat) over light-demanding plants. The food web is anchored by producers in the open areas and by dead matter under the tree.

Evaluation: five throws per area is a modest sample of patchy ground, and mobile animals violate the "count what's inside" rule by fleeing or wandering in. One survey is one moment — season and weather shift everything. Improvements: ten throws per area, a fixed survey route repeated across seasons, and pitfall traps (with daily checks and release) for shy invertebrates.

Safety and ethics

Fieldwork ethics are Leave No Trace applied to science: gloves on, unknown fungi and insects observed but not handled, hands washed, and every organism left alive in place — a quadrat survey should change nothing but your knowledge. Take care with sun exposure and any stinging plants or insects known on your grounds, and tell an adult where the group is working.

Common mistakes

• Placing the quadrat where it "looks good", destroying randomness and inflating the estimates.
• Reporting scaled-up estimates as exact counts ("there are 13,632 clover plants") instead of estimates with a sensible rounding.
• Drawing food-web arrows from predator to prey — arrows follow the energy, from food to feeder.
• Forgetting detritivores and litter, leaving the under-tree community unexplained.

Exam Experience

Question 1 — Analysis (4 marks). Using a 0.25 m² quadrat, a student finds a mean of 8.4 daisies per quadrat in a meadow of 500 m². Estimate the total daisy population. A second meadow survey gives 2.2 per quadrat — give the density per m² there. *Scoring guide:* Density: 8.4 × 4 = 33.6 per m² (1 mark). Total: 33.6 × 500 = 16,800 daisies (1 mark method, 1 mark answer, accept ≈ 17,000). Second meadow: 2.2 × 4 = 8.8 per m² (1 mark).

Question 2 — Evaluation and design (4 marks). A student's quadrat results vary wildly between throws in the same area (clover counts of 2, 18, 5, 21, 4). Explain what this variation means for the population estimate and give two ways to improve confidence in it. *Scoring guide:* High variation means the habitat is patchy and the mean (10.0) is uncertain — the estimate could be far off (1 mark). Improvements (1 mark each, max 2, plus 1 mark for the reasoning that more throws shrink the uncertainty of the mean): more throws; larger or multiple quadrats; random coordinates rather than throws; report the range alongside the mean.

Multiple choice 1. Quadrats must be placed randomly mainly to: A) save time B) avoid bias from choosing "interesting" spots C) count more species D) keep the area tidy. *Key: B.* Random placement makes the sample representative; chosen placement systematically over-samples rich patches and inflates estimates.

Multiple choice 2. In a food web, the arrow in "clover → snail" points from clover to snail because it shows: A) which organism is bigger B) the direction of energy flow — the snail eats the clover C) which was found first D) the direction the snail moves. *Key: B.* Arrows trace energy from food to feeder throughout ecology; reversing them is the classic food-web error.

A second look: counting what runs away — mark, release, recapture

Quadrats fail on animals that move. The ecologist's answer is mark–release–recapture: catch animals, mark them harmlessly, release them, then catch again and use the proportion marked to estimate the population. A class marked 24 snails with a dot of paint and released them. Next day they caught 30 snails, of which 9 were marked. The logic: the marked fraction in the second catch (9 ÷ 30 = 30%) should mirror the marked fraction in the whole population (24 ÷ N). So N = 24 × 30 ÷ 9 = 80 snails — the Lincoln index. Its assumptions are a checklist worth memorising: marks must persist and not affect survival or catchability, the population must be closed (no births, deaths or migration between catches), and marked animals must remix randomly.

Set the estimate beside the survey's snail density (1.6 per quadrat near the wall = 6.4 per m²) and the two methods interrogate each other: the index says how many snails the whole habitat holds; the quadrats say where they concentrate. Abundance, distribution, identity — a food web needs all three, and no single technique in this lesson supplies more than two.

Final check: when have you thrown enough quadrats?

Richness answers its own sampling question if you record it throw by throw. In open grass, the *cumulative* species count went: throw 1: 4 species, throw 2: 5, throw 3: 6, throw 4: 7, throw 5: 7 — the curve flattens: the fifth throw added nothing new. Under the tree: 2, 3, 4, 5, 5 — same flattening at five. A species accumulation curve that has levelled off certifies that your sampling effort was adequate for comparing these areas; one still climbing would order more throws before any conclusion is drawn. This is the design tool that converts "we did five throws because the worksheet said so" into "we did five throws, and here is the evidence five were enough".

It generalises beyond quadrats: germination counts, chromatogram repeats, jar surveys — every practical in this course contains a hidden accumulation curve where extra effort buys less and less new information. Skilled investigators spend their throws where the curve is still rising and stop where it flattens. That judgement — evidence about the evidence — is the real final experiment of the course, and you now have the tool for it.