Slime moulds

Fungi feed by absorption: they secrete enzymes into their surroundings and take up what dissolves. Slime moulds feed by engulfing — flowing around a bacterium and taking it inside, the way an amoeba does, because that is what they are. A fungal cell wall is made of chitin; a crawling plasmodium has no cell wall at all, which is how it changes shape.

The reason they were filed with fungi for so long is the only thing the two groups share: at the end of the growing phase, both raise small stalked structures that release spores. It is a good solution to the problem of getting spores into moving air, and it has been arrived at independently more than once.

A plasmodium grows the way most cells cannot: its nuclei divide, repeatedly and in near-perfect synchrony, while the cell itself never splits. The result is a single bag of cytoplasm that can spread over a log, with no internal partitions and no centre.

It moves by shuttle streaming. Cytoplasm surges one way along a tube, pauses, and surges back, with a rhythm of roughly a minute; where more streaming is needed the tube thickens, and where less is happening it is reabsorbed. There is no organising centre deciding which tubes matter. Thick tubes are simply the ones that were carrying more.

That mechanism is the whole explanation for the behaviour slime moulds are famous for, which is worth holding onto before reaching the next section.

The experiment is genuinely striking and the framing usually given to it is not what happened. The organism did not enter the maze and search it. It was spread through the entire maze first, occupying every corridor, and then food was placed at two exits. What it did next was withdraw from everywhere that was not on the path between them.

The mechanism is the streaming described above. Tubes on a route between two food sources carry more flow; tubes carrying more flow thicken; tubes carrying less are reabsorbed. Run that rule everywhere at once and short connections survive. Nothing represents the maze, nothing compares routes, and nothing chooses.

This is not a deflation. A system with no memory of the layout and no ability to see it still arrives at the short path, which tells you something useful about how little machinery a good solution requires. Calling it intelligence replaces that finding with a comparison to ourselves, and the comparison is the least informative part.

This result is stronger than the maze and much less famous, which is usually a sign that the famous one is being reported for its headline. Habituation has a formal definition with tests attached: the response must decline with repetition, the decline must be specific to the stimulus experienced, and it must recover after a rest. A tired or damaged organism fails those tests. A habituating one passes them.

What it shows is that learning, in its simplest form, does not require neurons. What it does not show is memory in the sense people usually mean — recollection of a particular event — or anything like understanding. The distinction matters because the popular version of this finding has already merged with the maze into a general claim about slime mould intelligence, which neither experiment supports.

A spore lands in something damp and releases a single amoeba, which may or may not grow a pair of flagella and swim. These cells feed on bacteria, divide normally, and behave in every respect like ordinary single-celled amoebae.

Two compatible cells then fuse. The resulting cell begins dividing its nucleus without dividing itself, and grows into a plasmodium. It feeds, spreads, and when food runs short or conditions dry it converts into spore-bearing structures — often within a few hours, and often after moving into the open where air can carry the spores away.

1. Spore germinates into an amoeba, sometimes flagellate.
2. Amoebae feed on bacteria and divide.
3. Two compatible cells fuse.
4. Nuclei divide without the cell dividing: a plasmodium.
5. Plasmodium feeds and spreads, streaming through litter or wood.
6. Conditions change; the plasmodium moves out and forms stalked spore capsules.

The history here has an unusual shape. The taxonomic correction and the behavioural results arrived from opposite directions and about a century apart, and the popular understanding has absorbed the second while largely ignoring the first — which is why "weird fungus that solves mazes" remains the standard description of an organism that is neither.

1. 1753

   First observation

   Slime moulds enter the record as fungi

   Linnaeus places the organisms that would become the myxomycetes among the fungi, on the reasonable grounds that they appear on rotting wood and make spores on stalks.

2. 1860

   Reinterpretation

   De Bary argues they are not fungi at all

   Anton de Bary watches the feeding stage move and engulf particles, concludes it is animal-like rather than fungal, and coins Mycetozoa — "fungus animals" — for the group.

   Changes how the 1753 result reads

   The evidence was behavioural and available to anyone with a microscope and patience. It took molecular data more than a century later to make the placement stick.

3. 2000

   Landmark experiment

   A plasmodium finds the short path through a maze

   Physarum polycephalum, spread through an agar maze and offered food at two exits, withdraws from the dead ends and keeps a tube along the shortest connecting route.

   Maze-solving by an amoeboid organism

4. 2005

   Replication

   Molecular phylogeny settles the placement

   Sequence data place the myxomycetes firmly within Amoebozoa, confirming de Bary and separating them from fungi by a distance comparable to that between fungi and animals.

   Changes how the 1860 result reads

   A rare case of a nineteenth-century inference from behaviour being confirmed almost exactly by twenty-first-century sequence data.

5. 2016

   Modern discovery

   Habituation demonstrated without a nervous system

   Plasmodia crossing a bitter but harmless bridge speed up over days, specifically for the substance they met, and forget after two days — the formal criteria for habituation.

   Changes how the 2000 result reads

   The stronger result and the less famous one. The maze can be explained entirely by flow reinforcement; habituation is tested against its own controls and cannot.

   Habituation in non-neural organisms: evidence from slime moulds

• Where is habituation stored in a cell with no nervous system?

  Why it matters: The behaviour meets the criteria and nobody has identified the cellular change that holds it. An answer would say something general about how minimal a memory can be.

• How closely does tube reinforcement match the optimisation rules used to model it?

  Why it matters: Physarum is used as a model for network design, from rail systems to algorithms, and the analogy is only as good as the match between the biology and the maths.

• How many times have stalked spore structures evolved in the amoebozoans?

  Why it matters: It is the trait that caused a century of misclassification, and knowing how often it arose would say how strong the pressure to solve that problem is.