Leeuwenhoek saw animalcules in the 1670s and bacteria from his teeth in 1683. For a long time after, microbes appeared in biology mainly as diseases and as recyclers. Both are true. Neither is the whole census.

The tree went through them

Carl Woese’s ribosomal RNA work did something ruder than add a chapter on bacteria. It made deep ancestry a sequence you could compare, and the comparison split prokaryotes into bacteria and archaea. Animals and plants sit inside a story whose trunk is microbial. The note on Woese and the species problem meet here: a great deal of microbial evolution is not a neat branching of species that keep their genes to themselves.

You are an ecosystem, in a limited sense

Skin, gut, mouth, and a few other surfaces host communities that help with some tasks and cause trouble in others. “Host” is already the careful word. The communities differ by person, by diet, by antibiotics, and by methods of counting. A sequencing read tells you which fragments were present in that sample. It does not, by itself, tell you which organism was alive, which was a passenger, or which gene was doing a job. Metagenomics is a remarkable census and a bad oracle.

Borrow the vocabulary and you can sell anything

Products advertise a “balanced microbiome” the way an earlier century advertised a balanced humour. Ecology does not have a single balance that a capsule restores. If a claim names a disease and a dose, it belongs in a clinical trial and a regulator’s file, not in this essay. The biology to keep is older and better: microbes cycle the elements, they made the oxygenated atmosphere possible, they live in places that would kill us, and a large fraction of them will not grow in our broth. That last fact is why Woese’s molecular census mattered. You cannot stain what you cannot see, and you cannot see the whole world through what happens to grow.