There's a bunch of fun simulation games you can play where you can sort of prove this to yourself. A given species doesn't just eat items X, Y, Z in a food web

There's a bunch of fun simulation games you can play where you can sort of prove this to yourself. A given species doesn't just eat items X, Y, Z in a food web. In a pinch, they might be able to deal with others things. A W, perhaps. If the W-predating species suffers disease or some calamity, this other species can, to some degree, pick up the slack. It's really a key part of adaptive radiation. Maybe I could live here, although I don't usually. Maybe I could mate there, although it is a bit out of my comfort zone. When you play those simulation games with just a few species, you see a lot of boom-bust cycles and eventually you get a collapse. Similarly, those sealed terrariums have similar problems: they eventually fall into a lower equilibrium of fewer species and then they kind of turn to mush. I hesitate to say "role" because that feels assigned, but different species do different things in an overlapping fashion, like a cross-trained staff lowering your "bus factor." Now, I imagine there's some kind of maximum number (so many species that finding someone to breed with is hard), but the minimums lack resiliency for various cycles.