Every student founds a population of 26 — species, color, pattern, tolerances, and how it raises its young — before anyone knows what the world looks like. Then a random ecosystem is drawn and the class population runs for 8–12 generations.
About the model: a teaching simulation, not a genome. Every design enters the world as 26 founders exactly as they were set; variation appears as they breed, because selection can only act on variation that is there. Populations of the same species sharing a world interbreed each generation, which is what holds a species together — and what the barrier takes away. Each lineage carries mean trait values that respond to selection by the breeder’s equation — the change per generation is heritability times the selection differential — under a fixed point budget, so every gain is paid for elsewhere. Color drifts toward whatever the substrate is because camouflage is what the predators are sorting on. Nothing here is tuned to make any one design win. Falconer, D. S., & Mackay, T. F. C. (1996). Introduction to quantitative genetics (4th ed.). Longman. · Lande, R., & Arnold, S. J. (1983). The measurement of selection on correlated characters. Evolution, 37(6), 1210–1226. · Endler, J. A. (1986). Natural selection in the wild. Princeton University Press. · Pianka, E. R. (1970). On r- and K-selection. The American Naturalist, 104(940), 592–597.