This week Sci continues her coverage of the IgNobel prizes, those prizes given for the greatest and WEIRDEST (but not at all useless) science on the planet. It's an award after Sci's own little heart.
And today we get to one of my personal favorites from this year:
GODZILLA SLIME MOLDS!!!
This is a wonderful paper because it's (a) extremely hilarious and (b) extremely INSPIRED. This experiment is a GREAT idea, and a very interesting way of looking at how we, as humans, interact with our environment. Perhaps we aren't so different from the molds after all.
Not only that, this is ANOTHER experiment that you might be able to do with your high school or college class, and one absolutely guaranteed to get some people interested.
Seriously though, I cannot understand how they got through this paper, AND the press coverage for this paper, without a Godzilla reference. NO FUN.
Why Godzilla, you ask? I answer: scale model of Tokyo.
But first let's talk about adaptive network design. Specifically, transport networks.
Transport networks (think planes, trains, automobiles, roads, streets, sidewalks, etc etc) are incredibly important to human society. The efficient movement of people, things, and information is what makes society what it is today. But for all that, most transport networks...aren't very well planned. Oh yes, the HIGHWAYS and stuff are well planned, but cities? Cities tend to end up springing up organically, mostly constrained by stuff that important at the time, but not really the most important in the grand scheme of transport.
Take London. London prior to 1666 still had its old Roman walls, but had grown up pretty disorganized:
The hit the Great Fire of London in 1666. It destroyed a large part of the city, and was generally really terrible. In the aftermath, with so much of the city gone, the King tried to get London redesigned, with piazzas and avenues and maybe a nice grid.
And we see how well that worked out. They did try, but there was a lot of confusion over property ownership, with people very unwilling to give up where their houses used to stand in favor of big wide avenues. They basically ended up recreating a lot of the old map, with improvements for wider streets and better sanitation.
So you have plans for efficiency, but these often get tossed over in favor of other constraints, like other people wanting their homes where they WERE, thank you.
Unfortunately, these networks don't often end up efficient, and can often fail (see baggage handling, airlines). So the scientists for this paper were interested in looking at network design, and interested in developing a model for how biological systems did networks, hoping to steal some of nature's ideas.
And they settled on the slime mold.
Now, let's talk about slime molds. I never knew it til now, but slime molds are FASCINATING creatures. They are often described as being a single humongous cell, but from what I understand, the stage where it looks like a mold and does stuff is actually a huge glom of cells, all joined together into one big cytoplasm with lots of nuclei. When it's all big like that, it has the ability to slime around and hunt for food. Slime molds eat microorganisms that feed on decaying plant and animal matter, so the usual place to find them is among the detritus of the forest floor.
But the cool thing is to watch them MOVE! They put out little teensy protoplasmic strands, and the cytoplasm rapidly follows. The whole slime mold can then stream along! Observe.
And these slime molds are smart (sort of). They can do MAZES!
Smart eh? That's from a thing that has NO BRAIN. NOTHING. A glom of cells! It's actually not smart, its just maximized its efficiency. They have a continuously expanding margin (easy to see in the maze video), with a tubular network extending out behind it, and lots of junctions to increase transport efficiency for already found food. The question was, how would the slime mold do when put in a situation where it has certain constraints? Where would its network form? And would its network represent the best possible efficiency?
Enter the map of Tokyo.
Well, not quite (much as it would have been TOTALLY COOL to use a full scale map to Tokyo). Instead, they used a map with the borders of Tokyo, and food sources placed at various small cities nearby. The slime mold was placed in Tokyo, and had to find the food sources. The idea was that, with the slime mold being highly efficient, the resulting network it came up with might be MORE efficient then what humans had come up with.
Here you can see in the top left the original map with the mold at Tokyo. As you go down through pictures B and C, the slime mold puts out its front, and follows behind, hitting each food point. When it gets to the edge, it draws the rest of the network back, ending up with a few filaments linking all the points on the map.
Very efficient. But how does this compare to the rail system that Tokyo has already in place?
A pretty good match, especially when they controlled for things like altitude and grade by using bright light (slime molds don't do light) to give it more constraints. In fact, the mold often did BETTER than the rail map, proposing new lines that added efficiency.
They even managed to model cost efficiency by looking at the possible number of links (high number is high cost). The mold did very well in terms of cost efficiency. The only part where it didn't beat the human system was in terms of fault tolerance, which is how well the system can adapt when something is cut off or fails. The human system of links had better fault tolerance, which may be because fault tolerance in this kind of environment (with lots of food nodes) may not be as necessary to the mold as it is to the humans.
But the really cool thing is this: based on the movement of the mold, the scientists were able to develop a MATHEMATICAL MODEL for expansion, creating a stable structure of connecting rail lines. This could be adaptable to all sorts of possible networks (though Sci thinks that using a slime mold to model the desperate need for decent Metro in the DC area would be a great start).
Of course, the only thing that could be cooler is if they'd used a real scale model of Tokyo. And dressed the mold up as Godzilla.
Tero, A., Takagi, S., Saigusa, T., Ito, K., Bebber, D., Fricker, M., Yumiki, K., Kobayashi, R., & Nakagaki, T. (2010). Rules for Biologically Inspired Adaptive Network Design Science, 327 (5964), 439-442 DOI: 10.1126/science.1177894