Community genetics investigates how genetic differences within a species affect the other organisms associated with it, and how those interactions can influence ecological processes and evolutionary change. Its gold standard is a clear chain of evidence from inherited variation to a trait, from that trait to an interaction, and from the interaction to a measured community outcome.
A tree's genetic background can matter to the insects, fungi and decomposers around it without determining every organism that arrives. This guide explains how to investigate such effects, distinguish genotype identity from genetic diversity, design informative comparisons and avoid treating a whole community as though it were a single organism. All numerical examples are original fictional teaching cases, not measurements from the studies cited.
Choose your community genetics route
Follow genetic variation through traits, species interactions, community measurements and environmental change.
Route 1 · Chapters 1–3: Define variation and mechanisms
Route 2 · Chapters 4–6: Design comparisons and measure communities
Route 3 · Chapters 7–9: Follow food webs, litter and inheritance
Route 4 · Chapters 10–11: Test feedback and environmental context
Route 5 · Chapters 12–14: Transfer, diagnose and continue
Chapter index
Define variation and mechanisms · Chapters 1–3
Design comparisons and measure communities · Chapters 4–6
Follow food webs, litter and inheritance · Chapters 7–9
Test feedback and environmental context · Chapters 10–11
Transfer, diagnose and continue · Chapters 12–14
An ecological community contains populations of different species occurring and interacting in a specified setting. Community genetics asks whether variation within one of those species helps explain differences in the wider assemblage. The focal species may provide food, shelter, chemical conditions or physical structure used by many others.
Imagine several trees of the same species supporting different insect assemblages. Those differences could arise from tree age, shade, location, previous damage, sampling effort or inherited traits. The genetic question is not answered merely by showing that the trees differ. It requires a design that can distinguish genetic background from these alternatives.
A genotype is a genetic constitution, defined at the resolution relevant to the study. A clonal lineage may allow researchers to replicate a genetic background across environments. A marker profile may help distinguish sampled individuals, but identifying different profiles does not by itself reveal which biological traits cause community differences.
The community outcome also needs a definition. Are we explaining the number of species, their relative abundance, the presence of a specialist herbivore, or a rate such as leaf-litter breakdown? These are related but distinct outcomes. A single number cannot stand in for every feature of a living assemblage.
Community genetics is therefore more specific than the general observation that genes and environments interact. Its distinctive move is to follow the consequences of genetic variation beyond the focal organism into the organisms around it. Molecular Ecology explains how molecular evidence serves ecological questions; community genetics asks which ecological consequences follow from genetic differences.
The scope here is ecological communities of organisms. The phrase community genetics is also used in some health-service contexts, but population screening and clinical service delivery are different subjects. Keeping that distinction explicit helps readers follow the biological question without confusing two uses of the same words.
A useful explanation has intermediate steps. A genetic difference may influence leaf chemistry, growth timing, tissue toughness or architecture. That trait can affect herbivore feeding or shelter. A change in herbivores may influence predators, and altered litter inputs may influence decomposers. Every additional step introduces a claim that needs evidence.
Consider the proposed chain genetic background → leaf toughness → herbivore performance → predator abundance. A correlation between genetic background and predator counts supports an association at the ends of the chain. It does not establish that leaf toughness is the mediator. Differences in leaf area or volatile cues might produce a similar pattern.
Original mechanism comparison
Suppose two replicated plant backgrounds, A and B, support averages of twelve and six herbivores per plant. A also has twice the leaf area. Per unit of leaf area, herbivore density could be identical. The apparent genetic effect on total abundance might operate through plant size rather than herbivore preference for leaf tissue.
Both interpretations can be biologically meaningful. A larger plant may genuinely support more organisms. But a claim about food quality requires a comparison that addresses the amount of available habitat. Decide whether the target is total community support or the effect per unit of habitat before choosing the denominator.
A stronger investigation measures the proposed trait, compares herbivore outcomes under controlled exposure and checks whether the trait difference persists across relevant environments. Manipulating one mediator can help, provided that the manipulation does not also change several other features without accounting for them.
Schweitzer and colleagues' cottonwood study connects genetically based variation in a tree trait with ecosystem processes. It provides a concrete research route through plant chemistry, litter and nutrient-related outcomes. The lesson is to measure the intervening biology, not to assume that every difference in DNA produces a large ecosystem effect.
A mechanism remains useful even when its effect is small. The question is how much it contributes under stated conditions, alongside the many other influences operating in the community.
Genotype identity asks which genetic backgrounds are present. Genotypic richness asks how many distinct backgrounds are represented. Evenness asks how equally they are represented. These quantities can change independently, and combining them under the single word diversity can hide the reason for an observed effect.
Original three-patch comparison
Imagine a plant species with three known genotypes, A, B and C. Patch One contains twelve A plants. Patch Two contains six A and six B. Patch Three contains eleven A and one B. The first has genotypic richness one; the other two have richness two. Yet their relative representation is very different.
A simple diversity measure is the probability that two independent draws with replacement have different genotype labels: one minus the sum of squared genotype frequencies. Patch One has value zero. Patch Two has value 1 − 0.5² − 0.5² = 0.5. Patch Three has value 1 − (11/12)² − (1/12)² = 22/144, about 0.153.
This calculation illustrates why richness alone is incomplete. It is not a measure of how genetically distant A and B are. Two labelled genotypes could differ at few or many relevant loci, and neither count tells us which traits differ.
Now replace B with C in Patch Two while retaining equal numbers. Richness and this evenness-sensitive measure remain unchanged. The associated insect community may nevertheless change if C has a different flowering time or defence profile. Genotype identity matters even when a diversity index is identical.
A sound experiment therefore uses multiple genotype combinations at each richness level. Otherwise, a contrast between one low-diversity mixture and one high-diversity mixture may really be a contrast between their particular members.
The practical question is precise: does the outcome depend on the number of genetic backgrounds, their relative abundance, their identities, or their interactions? The design should make at least some of those alternatives distinguishable before the result is described as a diversity effect.
CHAPTER 4 OF 14
4. Use common gardens without assuming they remove every confounder
A common garden grows different genetic backgrounds under shared, organised conditions. Replicated genetic backgrounds can help distinguish repeatable differences from variation caused by location. The garden is a comparison design, not a guarantee that every plant experiences exactly the same environment.
Light, soil moisture and neighbouring vegetation can vary across a plot. Plants may also carry different developmental histories, maternal effects or associated microbes. Randomisation and blocking help distribute some influences; repeated propagation or additional controls may be needed for others. The details determine which genetic interpretation is justified.
Original blocked design
Suppose four spatial blocks each contain one independently positioned replicate of genotype A and one of B. Herbivore counts for A across blocks are eight, twelve, ten and fourteen. Corresponding counts for B are five, nine, seven and eleven. Every within-block difference is three.
The raw counts rise and fall across the garden, but the paired comparison is consistent. This supports a repeatable A–B difference within these conditions. It does not yet identify the causal gene or establish that the difference persists in another climate.
Now imagine A occupies only the sunny half and B only the shaded half. An average difference of three could then reflect position. More plants within each half would improve precision about the two halves without resolving the genotype–location confounding. Replication cannot repair a comparison in which treatment and location are inseparable.
The independent unit matters too. Ten leaves sampled from one tree are not ten independently assigned trees. They can improve measurement of that tree, but they do not supply ten independent replicates of its genotype. Record both the number of trees and the number of within-tree samples.
The specialist guide to phenotypic plasticity explains why the same genetic background can express different traits in different environments. Community genetics extends that question: do those changes alter the associated organisms, and does the effect differ among backgrounds?
Species richness counts the species recorded. Abundance counts individuals or another defined quantity. Composition records which species occur and their relative representation. A community can change substantially while richness stays constant, so the response variable must match the claim.
Original composition example
Consider four insect species. Community A contains counts 8, 2, 0 and 0. Community B contains 2, 8, 0 and 0. Both contain ten individuals and two species. Their composition differs because the dominant species has changed. Community C contains 0, 0, 8 and 2: it has the same total and richness but no recorded species in common with A.
For equal sampling effort, a Bray–Curtis dissimilarity can be calculated as the sum of absolute count differences divided by the sum of all counts in both samples. Between A and B the numerator is twelve and denominator twenty, giving 0.6. Between A and C the value is one. These results express different degrees of compositional difference despite equal richness.
The calculation is descriptive. It does not identify a genetic cause. It also depends on sampling, abundance representation and the chosen dissimilarity measure. Communities with no observed organisms require special handling because a denominator can become zero; software defaults should not substitute for an explicit decision.
Detection matters. A larger or more complex plant may be harder to sample completely. An apparent reduction in insects could reflect lower capture probability. Standardise effort where appropriate and distinguish total abundance from abundance per sweep, per leaf area or per unit time.
Species identification also has an error structure. Preserve original observations and record later revisions rather than silently overwriting labels. The existing guide to reference identification sets and disagreement review develops this discipline for learning projects.
A useful report often needs several complementary outcomes. Show enough detail to explain what changed, while avoiding an indiscriminate search through indices until one happens to look convincing.
A mixture can perform differently from a single-genotype stand for several reasons. It may contain a particularly productive genotype, combine complementary traits, alter competition, or change the organisms that visit it. Before calling a result an interaction effect, define what the mixture would be expected to do without the proposed interaction.
Original additive expectation
Suppose equal-density stands of genotype A produce one hundred grams of plant biomass and stands of B produce sixty grams over the same interval. A simple equal-share additive expectation for a half-A, half-B mixture is eighty grams. This is a declared baseline, not a law that mixtures must obey.
If the observed mixture produces ninety-two grams, it exceeds that expectation by twelve grams, or fifteen per cent of eighty. It does not exceed the best monoculture, which produced one hundred. The statements above the average expectation and above the best component are therefore different claims.
Now consider a mixture producing seventy grams. It falls below the additive expectation but exceeds the B monoculture. Selecting only B as the comparator would make the mixture appear beneficial while concealing its shortfall relative to the stated baseline.
To interpret the mechanism, measure component contributions where possible. Did A become much more abundant, did both genotypes perform better, or did their relative proportions change? A total alone cannot distinguish those pathways. The expectation may also need adjustment if starting densities, survival or growth conditions differ.
Johnson, Lajeunesse and Agrawal's evening-primrose experiment investigated additive and interactive effects of plant genotypic diversity on arthropod communities and plant fitness. Its comparison of mixtures with component expectations illustrates why a simple diverse-versus-uniform contrast can leave important mechanisms unresolved.
Do not infer complementarity solely because a mixture has a larger total than one arbitrarily chosen genotype. Define the comparison, calculate it correctly and then investigate the processes that could explain the difference. Statistical departure from an expectation is a starting point for mechanism, not its complete identification.
CHAPTER 7 OF 14
7. Follow effects through food webs while keeping directions separate
A focal plant can influence herbivores directly and predators indirectly. Predators can then alter herbivore abundance or behaviour, changing the plant's experience in return. Community genetics becomes more informative when these directions are separated rather than compressed into a claim that the genotype controls the food web.
Original predator-access comparison
Imagine two plant genotypes tested with and without predator access. Without predators, average herbivore counts are twenty on A and twelve on B. With access, counts are eight on A and nine on B. Predator access is associated with a reduction of twelve on A and three on B.
The difference between those reductions is nine. Under a suitable randomised factorial design, this interaction contrast asks whether the effect of predator access depends on plant genotype. It does not by itself show that predators prefer A. A may differ in prey density, architecture or the ability of prey to hide.
An exclusion structure may also alter shade or humidity. A procedural control helps separate the physical effects of the structure from the removal of predators. If that control is missing, the interpretation must acknowledge the combined treatment.
Timing adds another complication. A count at the end of the experiment reflects arrivals, departures, reproduction and mortality. A lower herbivore count could follow predation, reduced colonisation or earlier resource depletion. Direct observations or repeated censuses can help distinguish these explanations.
A plant genetic effect on a predator is indirect when it operates through another organism or an altered environment. That does not make it weak or unimportant. It means the causal route includes a mediator that should be measured or tested.
The broader Evolutionary Ecology guide explains the difference between an ecological consequence and a closed evolutionary feedback. A food-web effect becomes part of an evolutionary explanation only when the relevant inherited variation and reproductive consequences are connected with evidence.
Genetic effects can continue after leaves fall. Differences in litter quantity or chemistry may influence decomposer activity, mass loss and nutrient dynamics. The relevant process then moves beyond the living plant while still potentially reflecting traits expressed during its growth.
Distinguish litter quality from litter amount. A tree may produce more slowly decomposing litter but contribute a larger total mass. The resulting ecosystem input cannot be predicted from a decomposition percentage alone. Both the material supplied and its subsequent transformation matter.
Original litter calculation
Suppose ten grams of dry litter from genotype A decline to six grams over a fixed interval, while ten grams from B decline to seven. The observed mass losses are forty and thirty per cent. This comparison holds starting mass constant, but it does not establish that all missing mass became a particular nutrient available to plants.
Now imagine whole-tree annual litter inputs of two hundred grams for A and three hundred for B. Applying the same illustrative loss fractions would give eighty and ninety grams lost respectively. B has the lower fractional loss but the larger total loss. The denominator changes the apparent ranking.
A stronger experiment can place litter from both backgrounds into several receiving environments. This separates litter origin from decomposition environment. If A litter loses more mass everywhere, the origin effect is consistent across those settings. If rankings reverse, the receiving environment modifies the relationship.
Litter bags also impose conditions. Mesh size can affect which organisms enter, and placement changes moisture or contact with soil. A useful interpretation describes the assay rather than treating it as a perfect miniature of unrestricted decomposition.
Nutrient concentration, total nutrient mass and net mineralisation are different measurements. A concentration can rise while total material declines, and mass loss can include leaching or fragmentation. Follow the specific quantity required by the ecological question.
The goal is an accountable material story: what entered, what remained, what moved and what was measured. Genetic background can contribute to that story without replacing the physical and microbial processes that perform the transformation.
CHAPTER 9 OF 14
9. Interpret community heritability with the population and design visible
Researchers can ask whether replicated host genetic backgrounds support repeatably different community measurements. Such analyses sometimes use the term community heritability. The term needs careful interpretation: it does not mean that every species in a community is inherited together as a fixed package.
Shuster and colleagues' cottonwood research analysed arthropod community variation associated with host genetic backgrounds and explored indirect genetic effects among species. This research makes an important distinction between genetic influences on community patterns and the stronger idea that whole communities reproduce as unified organisms.
Original variance illustration
Suppose a specified model attributes four squared units of variation in a community score to differences among replicated host genotypes and six to residual variation. The ratio four divided by ten is 0.4. Under the model's assumptions, this describes the relative contribution of the estimated genotype component in that population and setting.
It does not mean that forty per cent of any one insect is genetically caused by the host. Nor does it mean that forty per cent of the associated species will be passed to the next generation. The quantity concerns variation in a defined response among the sampled units.
If environmental variation increases the residual component from six to sixteen while the genotype component remains four, the ratio becomes 0.2. The genetic component did not disappear; its share of the total changed. Heritability-like ratios are conditional on the population, environment, measurement and statistical model.
Clonal replication can capture several forms of genetic difference together. Predicting response to selection in sexually reproducing populations requires attention to transmissible additive effects and genetic relationships. A broad genotype component should not automatically be inserted into an additive breeder's equation.
Quantitative Genetics develops variance components, inheritance and prediction. Here the additional challenge is that the measured outcome may summarise many interacting species. Its scale and biological interpretation need as much scrutiny as the genetic model itself.
Showing that host genotype affects a community establishes one direction of influence. A reciprocal evolutionary question asks whether the altered community changes reproductive differences among host genotypes. A complete feedback claim needs evidence for both directions, together with a plausible connection through time.
Original reciprocal-performance example
Imagine two host genotypes, A and B, tested in two experimentally characterised community conditions. In Community One, their expected successful offspring are twelve and eight. In Community Two, the values are six and ten. A has the advantage in the first setting; B has it in the second.
Starting with equal numbers of A and B in a simple asexual teaching model, the next A frequency would be twelve divided by twenty, or 0.6, under Community One. Under Community Two it would be six divided by sixteen, or 0.375. The community context changes the direction of selection in this model.
That result still does not establish a closed loop. We must also show that host composition helps generate the relevant community difference. If the two communities were supplied externally and never respond to the hosts, the experiment demonstrates community-dependent selection but not the whole feedback.
A time series can help establish sequence, yet sequence alone is insufficient. Weather might change both host performance and community composition. Controlled comparisons, mediator measurements or carefully justified causal models are needed to separate a feedback from a shared external driver.
Evolutionary change also requires transmission. A host's physiological response to its current community can alter reproduction without demonstrating inherited population change. The difference between plasticity and evolution remains important even when the ecological interaction is well established.
The practical discipline is to draw two arrows and label their evidence separately. One arrow runs from genetic background through traits to the community. The other runs from the community to relative reproductive contribution. If only one is supported, report that useful result without claiming the other.
A genotype can alter its phenotype across environments, and associated organisms can respond differently to the same phenotype under different conditions. Community effects may therefore change with season, moisture, temperature, nutrient supply or the available species pool. Transfer requires more than repeating the genotype label.
Original two-environment comparison
Suppose genotype A supports twelve herbivores per plant in a moist environment and four in a dry one. B supports eight in the moist environment and six in the dry one. The A-minus-B difference is plus four when moist and minus two when dry. The difference between those differences is six.
This is a crossing interaction in the fictional means. There is no single genotype ranking that describes both environments. Averaging equally across them gives A eight and B seven, but that average conceals a reversal that may matter for the question.
The effect can change even without a reversal. A might exceed B by four in one setting and one in another. Reporting only the overall average would then hide a change in strength rather than direction. Both patterns deserve attention when predicting beyond the measured environment.
Reusch and colleagues' seagrass field experiment examined genotypic diversity during extreme warming and found effects on plant and associated-fauna outcomes. The authors also discussed limits in separating recovery-related interpretations from productivity because experimental stands had not reached natural densities before the event. That qualification matters when using the study to reason about resilience.
A result from one disturbance, one species and one site is evidence about that system. It can motivate broader hypotheses without establishing that increasing genetic diversity always improves every ecosystem function.
For transfer, state the environmental range actually tested. Identify whether the proposed mechanism depends on a particular consumer, decomposer or resource. If that organism or condition is absent elsewhere, the same genetic difference may have a different ecological consequence.
CHAPTER 12 OF 14
12. Recalculate a conclusion when composition and environment change together
A useful transfer exercise changes more than the labels. Return to the moist and dry environments from the preceding chapter. In the moist setting, A supports twelve herbivores per plant and B eight. In the dry setting, the values are four and six.
Original mixed-stand forecast
Assume a stand contains one hundred plants, eighty A and twenty B, with no interactions among plants and the supplied per-plant means applying uniformly. In moist conditions, expected total herbivore abundance is eighty times twelve plus twenty times eight, giving 1,120. In dry conditions, the same composition gives eighty times four plus twenty times six, or 440.
Now change the stand to twenty A and eighty B. The moist prediction becomes 880, while the dry prediction becomes 560. Moving towards B decreases the predicted total in one environment and increases it in the other. A recommendation based solely on the moist comparison would not transfer.
Suppose the mixed stand also changes predator access, so the per-plant values no longer remain fixed. The additive calculation becomes an explicit null expectation rather than a complete forecast. Compare observations with it, then investigate which assumption failed. Do not preserve the original numbers while claiming to have incorporated the new interaction.
The target outcome matters too. More herbivores may increase food for predators while increasing plant damage. Neither total is automatically a universal measure of ecological quality. Define the objective and report trade-offs rather than selecting whichever outcome makes a preferred conclusion look strongest.
Finally, a stand with fewer surviving plants needs a new abundance calculation. Frequency changes alone cannot describe the total habitat remaining. Track both genotype proportions and total host abundance.
Successful transfer keeps the causal reasoning while rebuilding the quantities that changed. Explain the environmental shift, the composition shift, the assumptions about interactions and the outcome being valued. That explanation is more useful than a general slogan that diversity is good or that one genotype is best.
Does a genetic association identify a causal gene?
No. A marker can track ancestry or be associated with many linked differences. Replicated genotype effects establish a different level of evidence from identifying a causal variant and its mechanism. Keep marker, genetic background, trait and community outcome distinct in the explanation.
Do more genotypes always produce more species?
No. Effects depend on genotype identities, abundance, traits, environment and the organisms able to colonise. Some mixtures may alter composition without changing richness. A study showing an effect in one system supports that result and a hypothesis for further testing, not a universal monotonic rule.
Is a common garden sufficient to prove a genetic effect?
It improves comparisons but does not automatically eliminate maternal, developmental, microbial or positional differences. Ask how the material was propagated and assigned, which units were replicated and which influences remain confounded. The interpretation should match the actual controls.
Does community heritability mean a community is a superorganism?
No. Repeatable community differences associated with host genetic backgrounds do not by themselves establish that the entire assemblage is inherited, reproduces or has a single fitness. Stronger claims need their own definitions and evidence rather than borrowing support from a variance ratio.
Can a DNA survey replace community observation?
It can provide valuable evidence, but sequence detection, organism abundance and ecological activity are different quantities. Sampling, marker choice and reference databases affect the result. Use molecular measurements to answer an explicit question and connect them with the relevant ecological observations.
What should a null result mean?
It means the study did not establish the proposed effect at its resolution and under its conditions. Consider uncertainty, sample size, environmental variation and whether the chosen outcome captures the mechanism. Neither absence of significance nor a wide interval proves that genetic differences never matter.
What is the quickest repair for an overconfident explanation?
Name the focal genetic variation, the proposed trait mediator, the associated organisms and the measured outcome. Then identify one plausible non-genetic alternative. If the study cannot distinguish that alternative, narrow the claim before collecting more examples of the same association.
Begin with community description. Record which organisms are observed, the host or substrate, sampling effort, date and relevant environmental conditions. Preserve photographs or other original evidence where appropriate. A careful observation record is useful even when no genetic conclusion can yet be drawn.
Next practise the distinction between richness, evenness and identity using labelled tokens. Construct equal-richness mixtures with different proportions and different members. Calculate the simple diversity measure used earlier, then explain what it does not measure. This builds the habit of matching an index to a question.
For design practice, sketch a blocked common garden on paper. Allocate replicated host backgrounds across blocks, identify the independent unit and decide which community outcome will be measured. Add a plausible soil gradient and check whether the design can distinguish it from genotype effects. This exercise can expose confounding before any organisms are moved.
Crutsinger and colleagues' goldenrod experiment provides a primary example of manipulating genotypic diversity within a dominant plant species and measuring associated arthropod and productivity outcomes. Read it alongside the evening-primrose and seagrass studies to compare questions, designs and contexts rather than searching for one universal effect size.
A research-reading notebook can separate four things: what was manipulated, what was observed, what mechanism was tested and what remains uncertain. Add the comparator for every percentage. This makes papers with different outcomes easier to compare without flattening their distinctions.
Continue to Molecular Ecology for measurement and sampling, Quantitative Genetics for variation and inheritance, and Evolutionary Ecology for reciprocal ecological and evolutionary change.
The wider reading route returns to The Gold Standard of Education, How X Works and The Gold Standard master. The central achievement is a defensible explanation of how variation within one species changes life around it, together with a clear account of where that explanation may fail.