Sampling Techniques: Quadrats, Transects and Mark-Release-Recapture

Ecology is a quantitative science. Statements like "there are more daisies in the unmown half of the lawn than the mown half" carry no scientific weight until they are converted into measurements taken under a defined sampling protocol. This lesson anchors Required Practical 11 (RP11) — the AQA 7402 fieldwork investigation into the distribution of a species and the statistical analysis of the resulting data. It covers quadrat sampling, transect design, mark-release-recapture for mobile organisms, the Simpson diversity index, and the statistical machinery (chi-squared, briefly) used to make inferences from sampled data.

Spec mapping: This lesson sits in AQA 7402 Section 3.7.4 (populations in ecosystems) and underpins Required Practical 11 (investigation of the distribution of a species using quadrats and transects, with statistical testing). Cross-references to Section 3.7.2 (statistical analysis, chi-squared test) anchor the statistical machinery. Refer to the official AQA specification document for the exact wording of both sections and the CPAC (Common Practical Assessment Criteria) detail.

Connects to: Population ecology and Lincoln-index calculations (lesson 0 of this course); chi-squared statistical testing of categorical distribution data (course 8 lesson 0 The Chi-Squared Test); antimicrobial sensitivity testing as a parallel sampling problem (course 2 — comparable issues of replication, control, and statistical inference).

Key Definition: Sampling is the process of selecting and measuring a representative subset of a population so that the properties of the whole can be estimated. The validity of any ecological conclusion depends on whether the sampling protocol was unbiased, replicated, and statistically tractable.

---

Why We Sample Rather than Count

In all but the smallest, simplest systems, a complete census is impossible. A 100 m × 100 m woodland may contain a million invertebrates, tens of thousands of plants, and a handful of mammals — counting every individual is unfeasible. Sampling lets us estimate population abundance, density, distribution and diversity from a tractable number of measurements, provided three conditions are met:

1. The sample is representative — it captures the variability of the population, not just one corner of it.
2. The sample is unbiased — the selection method does not preferentially include or exclude particular individuals or sub-areas.
3. The sample is replicated — multiple samples allow estimation of variability and statistical inference.

A-Level misconception watch. Students often write "to make the result more accurate" when "accuracy" is not the relevant concept — repeated samples give greater precision (lower variance, narrower confidence intervals) but do not necessarily improve accuracy (closeness to the true value) if the sampling method is systematically biased. Use both terms correctly.

---

Random vs Systematic Sampling

Random Sampling

Random sampling selects sample locations independently of any feature of the habitat. Each potential location has an equal probability of being chosen. It is the default for assessing the abundance of a species across a uniform (or uniform-looking) habitat.

Method.

1. Lay out two perpendicular tape measures along the boundaries of the study area to create coordinate axes.
2. Generate pairs of random numbers using a random-number table, calculator function or smartphone app. (Avoid throwing the quadrat over the shoulder — humans are systematically biased, and the quadrat is more likely to land in clear areas.)
3. Place a quadrat at each random coordinate pair.
4. Record species presence, percentage cover, or individual counts within the quadrat.
5. Repeat until a sufficient number of quadrats has been recorded.

Strengths. Eliminates selection bias; statistically tractable; assumptions of subsequent significance tests are satisfied.

Weaknesses. Inefficient where the habitat is patchy — many quadrats may land in uninteresting parts of the habitat. Does not capture environmental gradients.

Systematic Sampling

Systematic sampling places samples at regular, predetermined intervals — along a transect, at the nodes of a grid, or at fixed coordinates relative to a habitat feature. It is the default for studying the distribution of species along an environmental gradient.

When to use systematic rather than random:

• The habitat has an obvious gradient (rocky shore zonation, woodland edge to interior, salt-marsh elevation series).
• The research question concerns change in composition with distance rather than overall abundance.
• You want every part of the gradient represented.

Strengths. Captures gradients efficiently; intuitive for distribution mapping.

Weaknesses. Risks bias if the systematic interval matches a periodic feature of the habitat (e.g. ploughing furrows); subsequent statistical tests may need to account for non-independence of adjacent samples.

---

Quadrat Sampling

A quadrat is a frame of known area used to define a sample plot.

Quadrat Types

Type	Description	Best for
Frame quadrat	A simple square frame, typically 0.25 m × 0.25 m or 1 m × 1 m	Counting individuals or estimating percentage cover
Gridded frame quadrat	A frame quadrat subdivided into 25 or 100 smaller squares by string or wire	Estimating percentage cover; ACFOR scale assessment
Point quadrat	A frame with vertical pins; species touching each pin are recorded	Vegetation analysis where layered cover is important
Open quadrat (corner pegs only)	Defines an area without a continuous frame	Larger sample areas, surveys of mobile or wide-spreading organisms

What to Measure in a Quadrat

• Density — number of individuals per unit area. Suitable for discrete, easily-counted organisms (e.g. snails, individual plants).
• Frequency — proportion of quadrats in which the species is present. Simple to record but coarse.
• Percentage cover — proportion of the quadrat surface occupied by the species (estimated visually, or via gridded quadrat). Suitable for plants where individuals are hard to delineate (grasses, mosses).
• ACFOR scale — Abundant / Common / Frequent / Occasional / Rare. Semi-quantitative; useful for rapid surveys but limits statistical analysis.

Quadrat Size Selection — the Greig-Smith Approach

Quadrats too small to capture multiple individuals of the target species give noisy, near-zero counts; quadrats too large are slow to sample and dilute fine-scale variation. The Greig-Smith curve (paraphrased classic approach) selects an appropriate quadrat size empirically:

1. Place a small quadrat at a random location.
2. Record the species (or number of individuals).
3. Double the quadrat area and record again.
4. Continue doubling.
5. Plot cumulative species count (or stabilised count of individuals) against quadrat area.
6. The "knee" of the curve — where the count stops increasing rapidly — gives a sensible minimum quadrat size for that habitat.

For a temperate grassland, quadrats are typically 0.25 m × 0.25 m to 1 m² for plant counts; for a rocky shore, 0.5 m² is common; for a woodland canopy survey, 10 m × 10 m or larger may be needed.

Sample Size

How many quadrats are enough? Three approaches:

1. Pragmatic — survey 10 to 30 quadrats and check that the running mean stabilises.
2. Power analysis — calculate the sample size needed to detect a stated effect at a stated significance level (beyond AQA spec, but the right approach).
3. Standard practice — most A-Level fieldwork uses 10–30 quadrats per treatment.

---

Transect Sampling

A transect is a line laid out across a habitat along which samples are taken. Transects are the standard method for studying distribution along a gradient.

Transect Types

flowchart LR
  A["Habitat A"] -- "Line transect" --> B["Habitat B"]
  C["Habitat A"] -- "Belt transect<br/>(continuous strip)" --> D["Habitat B"]
  E["Habitat A"] -- "Interrupted belt transect<br/>(quadrats at intervals)" --> F["Habitat B"]

Transect type	Method	Use
Line transect	Record every organism touching the line (or species present at fixed intervals)	Quick survey of zonation
Belt transect	Place quadrats continuously along the line	More detailed quantitative data
Interrupted belt transect	Place quadrats at fixed intervals (every 1 m, 5 m, 10 m)	Compromise between detail and efficiency — most common at A-Level

When to Use a Transect

• Rocky-shore zonation surveys (vertical gradient from upper-shore Pelvetia to sublittoral kelps).
• Salt-marsh elevation studies.
• Woodland edge-to-interior transitions.
• Studies of footpath impact on adjacent vegetation.

The independent variable is usually distance along the transect (or some abiotic factor that varies with distance — soil moisture, salinity, light intensity, pH); the dependent variables are species abundance or community composition.

Replication of Transects

A single transect captures one slice through the habitat. For statistical inference, multiple parallel transects (or multiple lines through different parts of the habitat) are essential. Reporting that "species X declined along the transect" without replication is a routine mark-loss.

---

Required Practical 11 — Investigation of the Distribution of a Species

This is the AQA 7402 RP11 anchor. The full protocol:

Aim. Investigate the distribution of a named species (typically a plant — daisies, plantain, lichen on tree trunks — or a sessile animal — barnacles, limpets on a rocky shore) in relation to an abiotic factor (light intensity, soil moisture, distance from path).

Method.

1. Define a clear research question — e.g. "Is the percentage cover of Trifolium repens (white clover) higher in regularly-mown lawn than in unmown grassland?"
2. State a null hypothesis: "There is no significant difference in percentage cover of T. repens between the two areas."
3. Design the sampling protocol:
   • Random sampling within each treatment area (mown vs unmown).
   • Sufficient replicates (typically 20 quadrats per treatment).
   • Standardised quadrat (e.g. 0.5 m × 0.5 m gridded).
   • Standardised measure (percentage cover read off the gridded subdivisions).
4. Record additional variables to characterise the habitat (light intensity with a light meter, soil moisture with a moisture probe).
5. Calculate mean percentage cover for each treatment.
6. Apply a statistical test — most commonly chi-squared on categorical-count data, or a t-test on continuous percentage-cover data after appropriate transformation.
7. Interpret in terms of biology — why might T. repens favour grazed swards? (Tolerance of low mowing; nitrogen-fixing root nodules; faster regrowth than competing grasses.)

CPAC criteria (the practical-endorsement assessment) reward: (i) following a written procedure; (ii) applying investigative approaches; (iii) safely using a range of practical equipment; (iv) making and recording observations; (v) researching, referencing and reporting.

Exam Tip. Specimen questions on RP11 typically ask either (i) to design a sampling protocol for a stated habitat and species, or (ii) to evaluate a flawed published method. Always argue from the principles — bias, representation, replication, independence — rather than from rote recall.

---

Mark-Release-Recapture and the Lincoln Index

For mobile species, quadrats are useless — the animal will not stay still long enough to be counted reliably. Mark-release-recapture (MRR) is the standard alternative; the technique was introduced in lesson 0 and is recapped and extended here.

The Lincoln Index

N = (M × C) / R

where:

• N = estimated population size.
• M = number of individuals captured, marked and released in the first sample.
• C = total number of individuals captured in the second sample.
• R = number of marked individuals recaptured in the second sample.

Assumptions (Six Conditions for Validity)

1. Closed population — no births, deaths, immigration or emigration between samples.
2. Marks are retained until recapture.
3. Marks do not affect survival or behaviour — marked individuals are not more visible to predators, more avoided by mates, or more or less trap-prone.
4. Random mixing of marked and unmarked individuals between samples.
5. Equal catchability — every individual has the same probability of capture in each sampling event.
6. The second sample is itself random with respect to the population.

If any assumption is violated, the estimate is biased. Direction of bias matters in mark-scheme answers — see the worked answer in lesson 0.

Worked Example

20 ground beetles are captured in pitfall traps, marked with a dab of nail-varnish on the elytra, and released. A week later, 25 beetles are caught, of which 4 are marked.

N = (20 × 25) / 4 = 125 ground beetles.

Refinements Beyond Spec

The Lincoln index is the simplest of a family of capture-recapture estimators. The Schnabel and Petersen-corrected estimators handle three or more capture sessions and small samples respectively. CMR (capture-mark-recapture) software (e.g. MARK) is used in research ecology. Beyond AQA, but useful at interview.

---

Simpson's Index of Diversity