A yellow network spreads across a surface, linking patches of food. Later, some tubes have thickened and others have disappeared. The route is not a map left by a traveling organism. It is the changing body of Physarum polycephalum itself, an organism that reorganizes how its material reaches nourishment.
Despite the name slime mold, Physarum is not a fungus. Its feeding plasmodium is a large cell containing many nuclei. It has no brain or neurons, yet can adjust to complicated spatial conditions. The interesting question is how local processes within a living body generate an organized result, rather than where a tiny planner might be hiding.
A maze experiment begins with a body already spread out
Toshiyuki Nakagaki and colleagues studied path finding by placing nutrients at two selected positions in a maze occupied by an extended plasmodium. The organism withdrew from dead ends and reorganized into a connecting tube along a short route. The setup matters: this was not simply a tiny creature standing at the entrance and mentally choosing its first turn.
Food locations changed which parts of the existing body were retained and reinforced. The final route emerged as other possibilities diminished. An endpoint photograph can make the process look like drawing a correct line through a puzzle. The experimental sequence shows something different: exploration and refinement through changes in living material.
That difference is useful when reading any maze headline. Ask where the organism started, what information it could encounter and what counted as success. Finding a route, retaining the shortest route and learning a route across repeated trials are different experimental questions. The word “solved” can hide those distinctions unless the method comes with it.
A transport network has more than one good property
In a 2010 study, Atsushi Tero and colleagues arranged food sources to represent urban locations around Tokyo and compared the resulting networks with rail infrastructure. The slime mold achieved a comparable balance among selected measures of cost, transport efficiency and tolerance of damage. It did not recreate every engineering or social requirement of a real railway.
A network's economy is not the same as the shortest journey through it. Connecting everything with very little material can create long detours. Adding an alternate connection costs material but can preserve access after a break. The study is interesting because those goals compete, leaving a range of useful compromises rather than one obviously perfect design.
Try the four-corner problem
Imagine four places at the corners of a square, labeled A, B, C and D clockwise. Allow links only along the sides, with every side costing one unit and taking one unit to traverse. Compare two designs: the chain A–B–C–D, and the same chain with an additional D–A link closing the loop.
Before reading the table, answer three questions. Which design costs less? What is the longest shortest trip between any pair of places? And can every place still reach every other after any one link breaks? Count trips along links, not straight through the square's interior. This is an invented puzzle with stated rules, not a reconstruction of the Tokyo experiment.
| Design | Total link cost | Longest shortest trip | All places connected after any one break? |
|---|---|---|---|
| Chain A–B–C–D | 3 units | 3 steps | No |
| Loop A–B–C–D–A | 4 units | 2 steps | Yes |
The loop spends one extra unit to improve the worst trip and provide a fallback route. After a break it becomes a chain, so journeys may lengthen even though everything remains connected. If construction material is scarce, the cheaper design may still have an advantage. The preferred answer depends on how strongly you value cost, distance and failure resistance.
Now imagine that food at D disappears. A fixed diagram remains unchanged unless someone edits it. A living network can redistribute material as conditions change. That continuing adjustment is one reason biological networks interest researchers: the construction process and later maintenance are parts of the same active system.
Flow and tube shape influence one another
The tubes contain moving cytoplasm. Rhythmic contractions drive internal flow, and their changing diameters affect how that flow is distributed. Nutrient encounters can alter contractions and tube architecture. Feedback between transport and structure helps explain how a response can become coordinated without an observer outside the organism directing every connection.
Mirna Kramar and Karen Alim investigated tube-diameter changes following a nutrient stimulus in 2021. They combined experiments with a model in which a locally released softening influence spreads through flows, favoring the enlargement of some tubes at the expense of others. Their interpretation described the resulting architecture as a form of memory of nutrient location.
That use of “memory” has been debated in a published critique. Demonstrated remodeling in response to nutrients is more specific than a claim of human-like recollection or consciousness. You can appreciate the experimentally observed relationship between food, flow and shape while keeping stronger cognitive interpretations as separate questions.
The solver is part of the solution
Oxford researcher Mark Fricker describes mathematical models that capture core features of these adaptive networks. A useful model identifies mechanisms sufficient to reproduce particular behavior; it need not reproduce every property of the organism. Comparing a model with observations can reveal which assumptions explain the network and which need revision.
The yellow tubes challenge a familiar expectation that organized problem solving must begin with a central representation of the whole task. Here, sensing, movement and the route itself are embodied in changing material. Understanding that process does not require deciding whether every successful adjustment deserves the word intelligence. It requires watching how the organism's next form grows from its current connections.
