This was a three day crash course in metacommunity theory. Not something I normally do, but I thought the idea of deep learning with tight feedback loops seemed great! (Credit: Helena Rosengarten) (Academic papers are not always helpful for learning context)

Lars suggested Hubbell, and I outsourced the decision of what to read after that to an LLM. Prompts I gave included: Identify pioneers in community ecology. For each, explain their main contributions and how their work shaped the field. Provide both historical and contemporary figures if applicable. Suggest two seminal papers for me to read that provide the most compelling defense of niche difference models and attack Hubbell's point of view. If you had to choose between reading ___ or ___ to learn core ideas in community ecology, which would you choose and why?

I only read for big ideas. There were lots of cool examples and details about studies that I could have taken notes on, but I expected I would get lost in the sauce (see metacognition vs. cognition, maybe even two types of cognition? Something like deep thematic connections vs. complete accounts of mechanisms).

I normally have trouble critiquing papers, but I didn’t struggle that much this time. Several things that helped:

• I kept dissertation ideas and other big ideas I’ve been thinking about in the back of my mind, so it was easy to identify salient bits of text. I was actually surprised by how much they jumped out at me!
• I paid more attention to background processes in my mind, e.g. if I was about to move onto another section but noticed an unresolved thought, I followed it. Often I only noticed the thought just as I was about to quash it, which felt very odd. Makes me think that those thoughts are always there but I don’t notice them
• I wrote down all the random ideas and questions that popped into my head while reading

Overall, I think it worked quite well and I’m excited to do it again soon!

Day 1: The Unified Neutral Theory of Biodiversity and Biogeography (Hubbell)

Niche assembly: communities are groups of species who presence, absence, and abundance can be determined by “assembly rules” based on function

Dispersal-assembly: communities are open, non-equilibrium assemblages of species thrown together by chance, history, and random dispersal

Dispersal assembly is considered neutral because all species are assumed to be equal in their probability of reproducing, moving, and dying

**It seems important to have a theory of how we get initial species presence and abundance—some sort of metapop model?

Hubbell’s theory shows that biodiversity equilibria will arise from speciation and extinction (in the same way as migration and extinction generate equilibria in island biogeography theory), leading to steady state distributions of species abundances that are correct on the timescale of millions of years

**I thought dispersal assembly and neutral models were supposed to be more realistic because they resist the temptation to equilibrate? Is the speciation vs. extinction equilibrium a fake equilibrium (nothing stays constant on such long timescales) or have I correctly identified some tension?

**Isn’t the mechanism for extinction in the neutral model niche difference and disturbance?

**What causes communities to leave their steady state?

If communities are accidental collisions, they are not highly coadapted

**Doesn’t recent research on phenotypic plasticity on rapid evolutionary timescales undermine this claim? Ecology and evolution can happen on the same time scale.

**Dispersal is a rapid/transient process, so it’s significant when the community’s starting from zero but seems less relevant after assembly. Niche differentiation and disturbance seem like the larger circuits. But niche differences influence dispersal events and dispersal interacts with niche differences. e.g. if you’re a bad disperser you won’t be featured in many dispersal events. But Hubbell’s response would be that bad dispersal strategies are trading off with some other trait that might be locally advantageous (heavy seeds?) so fitness differences equal out. But if this were true, no species would ever go extinct. So how do you get extinction in this model?