The ideas behind it
How did a slime mold plan a village?

At Thresh Architects, I set oats, the mold's food, on a 3D print of the land for a village we're planning, at the places I chose for the homes, gardens and gathering places, and in about a day the mold had grown paths between them, one tracing a riverwalk along the shoreline. The slime mold in the Tokyo study is one huge cell that spreads toward food as a web of tubes, then keeps the tubes linking the food it found and lets the rest fade. In a 2010 study, on a map of the Tokyo area with food at the cities, it grew a network similar to the area's railway (Tero and others).

In 2010, a slime mold grew a network like Tokyo's railway
The study was published in Science in 2010 by researchers in Japan and at Oxford (Tero and others). The kind of slime mold they used is one huge cell, the bright yellow blob you may have seen on mulch in summer, and it is harmless. Molds are fungi, and a guide from Utah State University says slime molds, despite the name, move and belong to a different kingdom of living things.
In a wet dish, on a small map of the Tokyo area (National Geographic), the researchers set food at 36 points marking its cities (Tero and others). Reports of the study say the food was oat flakes (Science News). The map was 17 centimeters (about 7 inches) wide, with the Pacific coast as its edge. They put a small piece of the mold at Tokyo, and it spread over much of the map, reached each piece of food, then thinned to a smaller set of larger tubes linking them.
The paper's pictures of one run end at 26 hours (Tero and others), and Science News says the network was built "after about a day." Because the mold avoids bright light, in some runs the researchers shone it on the parts of the map standing for mountains and lakes, which the railway has to go around.
On a map of the Tokyo area with oats at the cities, a slime mold grew a network similar to the area's railway in about a day.
The mold's network cost a little less than Tokyo's railway, which stayed connected better
Measured against the railway, the mold's networks scored about the same, a little cheaper and a little worse at staying connected. The paper concludes that its networks "showed characteristics similar to those of the rail network in terms of cost, transport efficiency, and fault tolerance" (Tero and others). Those are how much track or tube a network needs, how directly it links the cities, and how well it stays connected when a link breaks.
For cost, the paper measures each network's total length against the shortest set of links that could join every city. The railway's track was about 1.8 times that length, and the mold's tubes averaged 1.75 times as long over 21 runs (Tero and others). On the paper's measure of how directly each network links the cities, the two scored the same, and the mold reached that score at "marginally lower overall cost."
The railway did better at staying connected, which the paper calls "marginally better resilience" (Tero and others). Only 4% of breaks in a single link would cut any part of the railway off, against about 14% for the mold's networks grown with light for the mountains and lakes.
For the same transport efficiency, the mold's networks cost a little less than Tokyo's railway, and the railway was a little better at staying connected when one link broke.
The slime mold keeps the tubes that link its food and lets the rest fade

With nothing directing it, the mold first spreads out over the space around it, then keeps the tubes that link the food it has found and lets the others fade. The paper calls this way of growing "selective reinforcement of preferred routes and simultaneous removal of redundant connections" (Tero and others).
The researchers built a mathematical model of it, equations that copy how the mold grows, based on the flow of the mold's insides through each tube. In the words of the science writer Ed Yong, describing the model, "The faster the flow rate, the wider the tube becomes. If the flow slows, the tubes thin and eventually disappear" (National Geographic). Science News reports that within hours the mold itself began to refine its pattern, "strengthening the tunnels between oat flakes," while "the other links gradually disappeared."
With no plan, the mold spreads out over the space around it, then keeps the tubes that link its food and lets the others thin away.
For a village we're planning, I set oats on a 3D print of the land

The idea came from the Tokyo study, which I called "one of the weirdest science experiments I've ever seen" in a short video. I tried it on the land for one of the villages we're planning. First, I drew the existing conditions by hand, meaning the land as it is now, because it "slows me down enough to see the layers" of the land.
Then I made a 3D model of the land's shape on the computer and printed it, so I had a small copy of the ground with its hills and low places. I studied its contours, the lines that join points of the same height, "where water gathers, where it runs." The Tokyo study used light on a flat map and never tested a printed model of the land, so what the mold did on the print is what I saw in my own run.
I built a watertight chamber, tented it to keep the air humid, and started the slime mold. Utah State's guide says that a slime mold like the yellow blob, if it begins to dry out too quickly, forms a survival structure, a resting stage. The oats are the mold's food, so I set them on the nodes, the places people would go, "like homes, gardens, gathering places," and then let it grow.
I drew the land by hand, then set oats on a 3D print of it where the homes, gardens and gathering places would go.
The mold followed the shoreline, tracing a riverwalk

In about a day, and in a way I didn't expect, the mold followed the water line on the print, tracing where a riverwalk could go, a path along the water's edge. As I wrote in the script for the short, "At the ideal access points it reached the shoreline and pushed into the water." The ideal access points are the best places for people to get down to the water. Its tubes were "hugging contour, skirting ridges," meaning they ran along the slopes at one height and went around the ridges.
The mold also showed how far it would grow to reach the oats, which I took as "a clue that some nodes may see less foot traffic," meaning fewer people may walk to some of those places, and we'll need to give people more reason to visit them.
The Tokyo paper reports "a range of network solutions" across its repeated runs (Tero and others), meaning the runs grew somewhat different networks. So the riverwalk is what the mold grew on our print, and another run might grow something else.
On our print, the mold grew along the water's edge, tracing a riverwalk, and reached into the water at what I saw as the best places to get down to it.
You choose the places, and the mold finds paths between them
The mold grows toward the food it's given, so it finds ways between places someone has already chosen. I chose where the oats went, and reading what its network meant, such as which places may see fewer people, was my work too.
You don't need a slime mold to find those paths. Some planners, as I put it, "will let people wear down these paths before deciding where the sidewalks should go."
A path worn by people taking a shortcut across grass is called a desire path. A university news story reported that campus planners at colleges like Michigan State University left ground unpaved, let students wear paths across it, and then paved over them (UW-Madison News). Where should I build my house on my land? shows how a worn shortcut can guide the paths on your land.
I also want you to "try to break out of the standard grid," with everything laid out in straight lines at right angles. Then look for "the natural flow of a place," the routes people take across it.
The mold finds ways between the places it's given food, so the people planning the village still choose those places and read what its paths mean.
Sources
- Atsushi Tero and others, "Rules for Biologically Inspired Adaptive Network Design", Science, 2010
- Utah State University herbarium, "The Blob: Slime Molds", Fun Facts About Fungi
- Laura Sanders, "Slime mold is master network engineer", Science News, 2010
- Ed Yong, "Slime mould attacks simulates Tokyo rail network", National Geographic, 2010
- "Slime Mold Designed This Village", Thresh Architects on YouTube
- Parker Schorr, "Desire paths: the unofficial footpaths that frustrate, captivate campus planners", University of Wisconsin-Madison News, 2019
Let’s talk