Germs flow with water from lower to higher concentration

See image. Apologies for the infobox, I can't take another picture because I deconstructed the outflow when I saw what was happening.

The water supply at the bottom used to be clean. The plan was to have it flow slowly enough for germs to die off before they reach anywhere significant. However... the germs flow faster than the water. Not only that, they defy entropy by concentrating themselves in the last pool. Germs are entering the pool on the left, and germ concentration there is dropping. So is germ concentration in the middle pool. Note also that the mesh tiles don't matter: it happens with or without them.

This is counter to expected behaviour, because you would expect germs to behave mostly like ink: I'm not going to be able to clean a pool of water by running a more heavily contaminated stream over the top, yet that's exactly what's happening here.

This doesn't make it impossible to do passive cleaning, it just turns the design into something resembling nothing which would work in the real world: filtered sewage outflow pool, then immediately a drinking water intake pool, then a germ concentration pool behind that which gets pumped back to the sewage outflow pool. On the other hand, real-life designs (long tracks of interconnected pools in which germs slowly die off) will funnel all germs to the endpoint immediately.