This simulation models the foraging behavior of Physarum polycephalum, a real single-celled slime mold famous for solving mazes and approximating efficient transport networks — in a 2010 lab experiment it regrew a striking analogue of the Tokyo rail network when fed at city locations. Each agent here senses a simulated chemoattractant field (standing in for the organism's cAMP-like signaling) with left/center/right sensors and steers toward the strongest trail, depositing more pheromone as it goes. Real Physarum extends pseudopod-like protoplasmic tubes and can creep across a surface at up to roughly 1 cm per hour while hunting for food, and its externalized trail network acts as a primitive form of spatial memory. The reinforcing feedback between trail-following and trail-laying is what causes the loose cloud of agents to self-organize into sharp, branching, vein-like networks. Drag to orbit, scroll to zoom.