Guide: Biomorphs

Use the app as a short model-based investigation or a longer artificial-selection design challenge.

Learning Objectives

Distinguish genotype from phenotype in a computational model.
Explain how small inherited variations can accumulate over many generations.
Compare artificial selection by a human chooser with natural selection by environmental survival or reproduction.
Use evidence from a selection history to support a claim about cumulative change.
Identify model assumptions and limits, including symmetry, high mutation rate, and human-defined success.

Observe the parent. Record the generation number, Gene 9 branch depth, and one visible feature of the phenotype.
Read the litter. Pick one child and use the label to identify which gene changed and whether it increased or decreased.
Predict before selecting. Write one sentence predicting what will happen if that child becomes the next parent.
Select and compare. Choose the child, then compare the new parent with the previous parent.
Repeat. Continue for at least 10 generations, keeping the same selection goal.
Export evidence. Save the final image and configuration. Use the selection history to explain how small changes accumulated.

Teaching Move	Prompt for Students	Concept Target
Separate mutation from selection.	Which change appeared without regard to your goal, and which later choice favored a goal?	Random origin of variation; nonrandom selection among variants.
Track cumulative change.	How would the result differ if you could choose only from the first litter and never continue the lineage?	Cumulative selection versus single-step search.
Distinguish body and instructions.	Did the adult drawing reproduce, or did inherited values generate a new body?	Phenotype selection with genotype inheritance.
Name the selector.	Was the selector a person, a scoring rule, or an environment with consequences for reproduction?	Artificial selection and limits of the model.
Ask what is missing.	What would need to be added to study drift, gene flow, or changing population frequencies?	Natural selection is not synonymous with all evolution.

Launch question, 3 minutes: If a complex form appears after many small changes, does that mean the final form was planned at the beginning?
Model orientation, 5 minutes: Identify genes, parent, 3x3 litter, mutation legend, and selection history. Emphasize that genes are copied and bodies are regenerated.
Selection run, 10 minutes: Students pursue one visible goal, such as taller, wider, compact, or appendage-rich.
Pair comparison, 7 minutes: Students compare final images and histories. They identify two mutations that mattered and one mutation whose effect was surprising.
Exit explanation, 5 minutes: Students answer: What was inherited? What was selected? What did the model leave out?

Warm-up, 5 minutes: Students sketch what they think one mutation can do to a branching body.
Manual selection, 12 minutes: Groups start from the same ancestor but use different human goals.
Automated selection, 12 minutes: Groups rerun from the same ancestor using a named evaluator, such as Higher Shannon entropy or Tallness.
Evidence table, 8 minutes: Groups compare generation count, final image, changed genes, and evaluator score where used.
Model critique, 8 minutes: Students mark each model feature as realistic, simplified, or artificial.
Claim writing, 5 minutes: Students write a claim about cumulative selection using one exported image and one history entry as evidence.

One-gene comparison: Start from the basic tree. Inspect the mutation legend and choose variants that change only Gene 5. Describe the phenotype difference without referring only to the number.
Cumulative selection: Pick a goal such as "wide and low" or "insect-like." Select for 12 to 20 generations. Export the final image and configuration.
Parallel histories: Give all groups the same starting configuration but different selection goals. Compare final phenotypes and selection histories.
Development replay: Use Show Development and ask students to distinguish the order of drawing from inheritance. The animation shows development of one body, not evolution across generations.
Evaluator comparison: Run the same number of automated steps with two different evaluators. Compare what each scoring rule rewards.
Model critique: Ask which parts resemble biological evolution and which parts are artificial conveniences for classroom speed.

Did the selected body get copied directly into the next generation, or were genes copied and redeveloped?
Which parts of the process were random, and which parts were directed by a selection rule?
Which gene changes were easy to interpret? Which depended on the values of other genes?
Why does repeated selection over many generations differ from waiting for one perfect random variant?
Could two different histories arrive at similar-looking phenotypes?
What would have to change to model natural selection instead of artificial selection?
In real populations, why is fitness better understood as reproductive contribution than as simple survival?
How does the high mutation rate help classroom observation while making the model less biologically realistic?
How is an automated evaluator different from a real environment?
When a gene changes, why can several parts of the phenotype change at once?

Have learners submit a final configuration, image, and short explanation of three selected mutations.
Ask learners to annotate one row of the mutation legend and predict the next generation before selecting.
Use two exported configurations as evidence and ask which one likely experienced selection for width, height, or branching depth.
Ask learners to compare a manual run and an automated run, then identify which one has a clearer selection criterion.
Have learners revise a misconception statement such as "the body copied itself into the next generation."

Criterion	Strong Evidence	Developing Evidence
Genotype and phenotype	Explains that genes are inherited and phenotype is regenerated through development.	Identifies genes and body shape but treats them as direct copies.
Cumulative selection	Uses multiple history entries to explain change over generations.	Describes only the final phenotype.
Use of evidence	Connects exported image, gene values, and mutation legend to a claim.	Provides observations without connecting them to a claim.
Model critique	Clearly distinguishes artificial selection, natural selection, and model simplifications.	States that the model is simplified but does not explain how.

The app does not show organisms trying to change. The chooser selects among available variants.
The app is not a complete model of natural selection. It is mainly artificial selection on visible phenotype.
Natural selection is not the same thing as evolution as a whole. Other mechanisms can change populations, but they are outside this model.
"Fittest" does not simply mean strongest, fastest, or longest-lived. In real evolution, fitness is tied to reproductive success in a particular environment.
A gene value does not have a single isolated meaning in all backgrounds; genes can interact in development.
A visually complex phenotype is not drawn by hand. It is generated by repeated developmental rules.
The automated evaluator is not "nature." It is a visible rule that helps learners inspect how different selection criteria change outcomes.